Four new combinations in Chomelia and Stenostomum ( Rubiaceae , Guettardeae ) from Central America , the Guianas and the Amazon Basin

According to current generic delimitations, supported by wood anatomy, palynology, and recent molecular phylogenies, Antirhea, traditionally treated as a Pantropical genus, is instead restricted to the Paleotropics. At the same time, the Neotropical species traditionally positioned in Antirhea should be positioned in Stenostomum. Following these definitions, Stenostomum is a genus with most of the species present in the Caribbean Region, and two species present in the Guianas, namely S. acreanum and S. guianensis, distributed throughout the Amazon Basin. In addition, according to these new generic delimitations, two species previously positioned in Guettarda and Antirhea need to be transferred to Chomelia, as C. triflora and C. ulei.


INTRODUCTION
The generic delimitations within the tribe Guettardeae (subfamily Cinchonoideae) have been debated for more than two centuries.Guettarda has been traditionally recognized as a Pantropical genus of about 140 species, with the main centres of diversity in the South Pacific and in the Caribbean Region (c.80 spp.), and about 50 species in South America.Borhidi & Fernández (1995a, b), following the studies of Chaw & Darwin (1992, 1993) separated the Neotropical species usually included in Antirhea Juss., resurrecting the genus Stenostomum C.F.Gaertn., and reducing Antirhea as a genus restricted to the Paleotropics.According to Chaw & Darwin (1992), Borhidi & Fernández (1995b) and Borhidi (2006), Stenostomum is characterized by the 4-or 5-merous flowers, truncate or shallowly lobed calyx, corolla tube completely glabrous or sericeous outside, glabrous or pubescent at upper portion inside, and 3colporate, tectate, and 'punctitegillate' pollen grains; however, according to Achille et al. (2006) the pollen of Stenostomum lucidum (Sw.)C.F.Gaertn.has reticulate tectum (not punctite gillate).In addition, Borhidi & Fernández (1995b) divided Stenostomum into three sections: 1) Sect.Stenostomum, with resinous leaves, free stipules and 2-or 3-branched stigma; 2) Sect.Resinanthus Borhidi, with scantily resinous leaves, free or shallowly connate stipules and 2-or 3-branched style; and 3) Sect.Neolaugeria (Nicolson) Borhidi, with resinous leaves, connate stipules (forming a truncate ring) and 4-or 5-branched stigma.The last two sections were later demonstrated to be paraphyletic by recent molecular phylogenies by Moynihan & Watson (2001); based on their phylogenies, they returned the sect.Neolaugeria to the generic level, and only recognized three species restricted to the Antilles, corresponding to the same generic and specific delimitations proposed by Nicolson (1979).Achille et al. (2006) confirmed the distinctness of Neo laugeria from Stenostomum and further suggested that sections Stenostomum and Resinanthus together do not form a monophyletic group.Therefore, Stenostomum is here circumscribed corresponding to Stenostomum sect.Stenostomum; while sect.Neolaugeria is here maintained at the generic level (Moynihan & Watson 2001, Achille et al. 2006); and sect.Resinanthus is tentatively recognized as a possible separate genus (closely related to Chomelia Jacq.), pending future phylogenetic studies.Stenostomum, as here delimited, is a genus of about 30 species, mostly centred in the Caribbean Basin, and a few species in South America (see below).Some previous reports of occurrence in northern South America were published by Darwin (1979) and Gentry (1993); however, these reports might be referring to the genus Pittoniotis Griseb., treated by these authors as a synonym of Stenostomum.Pittoniotis, as delimited by Steyermark (1974: 863-867, f. 138), and the present authors, is a monotypic genus restricted to the Caribbean coast of Panama, Colombia and Venezuela, easily distinguishable from Guettarda, Chomelia, Stenostomum and Neolaugeria by the stamens wellexserted above the corolla, with filaments as long as the anthers (vs anthers sessile to subsessile), and the paniculate, many times branched inflorescence (vs 1-fewflowered or cymose).
Chomelia, as traditionally recognized, is a Neotropical genus of about 60 species of shrubs and trees, with the main centres of diversity in the Andean Cloud forests, the Guayana Highlands and the Brazilian Atlantic forest.In addition, Chomelia has been classically treated as the sister genus of Guettarda L., distin-guished by the persistent, lobed calyx (vs caducous, truncate in Guettarda), corolla lobes valvate, valvate-induplicate or rarely narrowly imbricate (vs imbricate in Guettarda), and fruits usually 2-locular (vs (1-2) 3-9-locular in Guettarda; Hooker 1873, Müller Argoviensis 1881, Bremekamp 1934, Steyermark 1972, 1974).However, the molecular phylogenies produced by Achille et al. (2006), showed that Chomelia is more closely related to Stenostomum sect.Resinanthus (represented in the study only by S. acutatum DC. and S. myrtifolium Griseb.),than to Guettarda.Of the genus Chomelia, only C. spinosa Jacq.was included, which provided preliminary evidence that Chomelia and Stenostomum sect.Resinanthus are sister taxa; nevertheless, because only one species of Chomelia was investigated, the monophyly of the genus was not significantly tested.In the same phylogenetic study, Guettarda was shown to be polyphyletic, and represented by several separated clades in the Paleotropics and in the Neotropics.Following these results, most paleotropical species of Guettarda were transferred to other genera (Achille 2006, Achille & Mouly in prep.).Therefore, much work remains to be done in order to detect monophyletic groups in the tribe Guettardeae.
A set of diagnostic characters for the separation of Chomelia, Guettarda, Stenostomum and Neolaugeria is presented in Table 1.According to this set of characters, three species previously positioned in Antirhea or Guettarda and occurring in the Guianas (one of them extending throughout the Amazon Basin) need to be transferred to Chomelia or Stenostomum, and the necessary new combinations are presented below.These combinations are also needed for the ongoing Rubiaceae treatment for the Flora of the Guianas (Delprete submitted).9 Feb. 1991 (fl).
Also, the vestiture of the hypanthium was described to be densely velutinous, but the hypanthium of Poncy 889 is densely white-hirsute.This difference in vestiture is here interpreted as a simple morphological variation within the species, and an emended description is provided above.

Chomelia ulei
-Fig. 2 Guettarda ulei K.There has been a considerable confusion regarding the identity of Guettarda ulei and G. acreana K.Krause, caused by either a mixed collection or a confusion of specimen labels.The holotype specimen of G. ulei was cited by Krause (1914) as Ule 9860, and probably existed at B. However, when Macbride took photos of the types preserved at B, the specimen labelled as the type of G. ulei, negative N. 388, was instead Ule 9707, which is the holotype of G. acreana.Standley (1936: 119)  redordii the calyx plus the hypanthium is 5-6 mm long, and the hypanthium is tubular-campanulate, densely pilose with long, adpressed, white hairs, while in C. ulei the hypanthium plus calyx is 1.5-1.8mm long, and the hypanthium is obovoid and sparsely pubescent.For these reasons, we prefer to treat C. ulei as a separate species, pending future comparative studies with all Amazonian species of this genus.
Both specimens of Ule 9709 at U and L, have the label "Guettarda Ulei Krause n. sp., Bl. grünlich gelb, Bm. ad Seringal S. Francisco, Rio Acre, September 1911" handwritten by Ernst Ule.Later, another hand (of unknown identity) struck out "ulei Krause n. sp." and wrote "acreana Krause in Notizbl.Berlin VI, 56 (15.V.1914) p. 204."This might explain the confusion between G. acreana and G. ulei.However, we prefer to choose the U specimen as the lectotype of G. acreana because it is a more complete specimen.
The position of G. acreana has been regarded as problematic by many authors.Standley (1934), who was familiar with the numerous Antirhea species from the Caribbean area, already suggested the positioning of this species within the genus when he described A. panamensis, here treated as one of the synonyms of this species.Similarly, Bremekamp (1959) re-described this species under the name A. surinamensis.In the same work, he also suggested that a worldwide study of Antirhea s.l."will lead to a splitting of the genus," which was supported by later studies.
Distribution & Ecology -Only known from the type collection.
When Steyermark (1972) described Guettarda fanshawei Steyerm., apparently he was not aware that this species was already described by Bremekamp (1952), using a duplicate of the same collection, as Antirhea guianensis Bremek.Therefore, the two taxa are synonymous, and a new combination in Stenostomum is necessary, using Bremekamp's basionym.The specimen preserved at NY has a label handwritten by Borhidi with the new combination in Stenostomum, but apparently he did not publish it.In addition, Steyermark (1972) already suggested a close relationship of G. fanshawei with G. acreana K.Krause (= Stenostomum acreanum, see above).
contributed to maintain this confusion by citing Macbride's negative N. 388 as "Ule 9860" and treating it as G. ulei, which is instead Ule 9707 and is the holotype of G. acreana.In the same work (p.122) he cited s.n.(GDC), Pohl s.n.(GDC)), as suggested by Standley.The two taxa are very similar by having subsessile, subcapitate, cymose inflorescence, calyx minutely lobed, corolla pubescent outside, 11-12.5 mm long, glabrous inside, and anthers presented among the corolla lobes.Krause described G. ulei as with corollas 16-18 mm long, but the corollas of the sole type specimen available are only 11-12.5 mm long.The only clear difference among the two taxa is in the calyx lobes, which are round, 4 mm long and obtuse at apex in C. sessilis, while in G. ulei are linear, 0.5-0.8mm long and acute at apex.A comparison of the fruits of the two taxa is impossible, as none are available in G. ulei.Also, in the protologue of G. ulei, the stigma is described as capitate, but in the specimen Ule 9860 (K) it has two flattened, oblong branches, as it is common in Chomelia.The most distinctive characters of this species are the sessile, condensed, several-flowered inflorescences, with many conspicuous bracts, that are also found in C. recordii Standl.(from Mesoamerica and Colombia), which are uncommon in Chomelia; however in C.