Curtia ayangannae, a new speCies of gentianaCeae from the pakaraima mountains, guyana

A taxonomic treatment of the Guianan Gentianaceae is currently being conducted by an international group of botanists for inclusion in the Flora of the Guianas. While working on the genus Curtia Cham. & schltdl. and related genera, the authors came across two collections that were clearly different from the already described species with regard to macromorphological characteristics. The growth habit, strongly ridged internodes, stem/branch colour, leaf characteristics, and floral shape and colour distinguished these collections from the other species (Fig. 1). An earlier study at the micromorphological level (Bouman et al., 2002) showed that the seeds also differ from the rest: they are much larger in size (at least 3–4 times), are laterally flattened, and have extended micropylar and chalazal ends. It was evident, therefore, that these two collections comprised a new species. one problem that did arise, however, was the genus – does the new species belong to Curtia or to one of the related genera, Tapeinostemon Benth. or Saccifolium Maguire & Pires? not only do the specimens share some characteristics with each genus separately (Maguire & Pires, 1978; Grothe, 1981; Maguire, 1981; simonis, 1984; struwe et al., 1999, 2002), but they also have characteristics that are found in two or even all three genera (Table 1). As noted below, the pollen of this species fits in both Curtia and Tapeinostemon (Fig. 2). Using the data at our disposal, the specimens seem to fit best in the genus Curtia. They have, among other things, the same kind of inflorescence (although dense), calycine colleters (absent in Tapeinostemon), no obvious connective (present in both Tapeino­ stemon and Saccifolium), no appendages present on the anthers/filaments (present in

A taxonomic treatment of the Guianan Gentianaceae is currently being conducted by an international group of botanists for inclusion in the Flora of the Guianas. While working on the genus Curtia Cham. & schltdl. and related genera, the authors came across two collections that were clearly different from the already described species with regard to macromorphological characteristics. The growth habit, strongly ridged internodes, stem/branch colour, leaf characteristics, and floral shape and colour distinguished these collections from the other species (Fig. 1). An earlier study at the micromorphological level (Bouman et al., 2002) showed that the seeds also differ from the rest: they are much larger in size (at least 3-4 times), are laterally flattened, and have extended micropylar and chalazal ends. It was evident, therefore, that these two collections comprised a new species. one problem that did arise, however, was the genus -does the new species belong to Curtia or to one of the related genera, Tapeinostemon Benth. or Saccifolium Maguire & Pires? not only do the specimens share some characteristics with each genus separately (Maguire & Pires, 1978;Grothe, 1981;Maguire, 1981;simonis, 1984;struwe et al., 1999, 2002, but they also have characteristics that are found in two or even all three genera (Table 1). As noted below, the pollen of this species fits in both Curtia and Tapeinostemon (Fig. 2).
Using the data at our disposal, the specimens seem to fit best in the genus Curtia. They have, among other things, the same kind of inflorescence (although dense), calycine colleters (absent in Tapeinostemon), no obvious connective (present in both Tapeino stemon and Saccifolium), no appendages present on the anthers/filaments (present in 1) nationaal Herbarium nederland, Utrecht University branch, Heidelberglaan 2, 3584 Cs Utrecht, The netherlands. 2) nationaal Herbarium nederland, Universiteit Leiden branch, Einsteinweg 2, 2333 CC Leiden, The netherlands.
Tapeinostemon and Saccifolium), and the same leaf shape (suborbicular leaves: C. ob tusifolia and C. rotundifolia have rotund/obtuse leaves). The specimens also share some characteristics with Tapeinostemon. They are woody (shrub-like; Curtia species are annuals), grow in the same habitats (tepuis), and have corolla lobes that are shorter than the corolla tube, a prominent leaf venation, and a filament insertion in the middle of the corolla tube (this is variable in Curtia). some of the species of both Curtia and Tapeino stemon are heterostylous, as is the monospecific genus Saccifolium (Maguire & Pires, 1978;Grothe, 1981;simonis, 1984;struwe et al., 2002;struwe, 2006). Because of the small number of specimens available, it is unknown whether this new species is also heterostylous. The main characteristics the two collections share with the genus Saccifolium are probably caused by adaptation to their habitat. The specimens were collected on Mt Ayanganna at altitudes between 1800 and 2040 m. They were found growing in depressions (tepui bogs) on sandstone together with Bonnetia, Brocchinia, Clusia, Eriocaulaceae, Podocarpus, Rapateaceae, and Xyridaceae. Like Saccifolium, they have a thick corky bark and leathery leaves with strongly revolute margins (saccate in Saccifolium) that are crowded at the branch apices. The wings/ridges on each internode, however, are very characteristic: they pair off decussately, with each pair forming a 'cushion' at the top of the internode from which the petiole arises. The wings/ridges of each pair then taper towards each other and sometimes fuse at the base. Because of their overlapping characteristics, the genera Curtia and Tapeinostemon are difficult, if not impossible, to key out in the field. The authors feel, therefore, that a thorough study of the two genera is necessary. not only should more specimens be collected and wood and leaf samples (including this new species) investigated, but studies at the molecular level need to be performed. Perhaps these two genera should be combined into one genus. If so, the genus would retain the oldest name, Curtia. shrub, 0.25-0.5 m tall with silvery, corky bark, woody at base, branched. Branches to 0.6 cm diam., semiterete, dark purple, rigidly 4-winged (wing-like where leaves are still attached, more ridge-like where leaves have fallen off), pairs of wings/ridges each form a horizontal ridge at the top of each internode from which the petiole grows and taper towards each other sometimes fusing at the base of each internode. Internodes 2-10 mm long, shorter at base, stem/branches appear segmented because of the break in the bark at each node. Leaves decussate, originally arranged along stem, clustered towards tip of branches at maturity; petioles 0.1-0.5 cm long, flattened, arising from top of ridge; blades subcoriaceous, suborbicular, (0.7-)1-1.5 by (0.5-)0.8-1.2 cm, longer than internode, margin revolute, apex obtuse, base obtuse to acute, attenuate, tripliveined, veins very distinct, impressed above, prominent beneath, continuing into petiole. Inflorescence dense panicles, terminal and in upper leaf axils, 1-3 cm long including 0.3-0.6 cm long peduncle; bracts and bracteoles linear, 1-2(-5) mm long (basal bracts often leaf-like, to 5 by 2 mm); pedicels (0.8-)1-2(-4) mm long. Flowers actinomorphic, 5-merous, calyx dark purple, 1.5-2.5 mm long, lobes narrowly triangular, c. 1.6 by 0.6 mm, somewhat scarious, apex acute, connate 1/5-1/4 from base, small glands (colleters) at base on inside; corolla urceolate/campanulate, dark purple to orange, 3.5-4 mm long, tube 2.6-3 mm long (longer than calyx lobes), lobes broadly ovate, c. 1 by 0.8-1.2 mm, papillose, apex acute, sometimes somewhat recurved at maturity, connate c. 3/4 from base; stamens c. 1.5 mm long, filament c. 1 mm long, straight, equal in length, inserted ± halfway from tube base, anthers broadly ellipsoid after anthesis, c. 0.5 mm long, pollen in monads, oblate spheroidal, 3-colporate; gynoecium c. 2.5 mm long, ovary narrowly ovoid, 1.8-2 mm long, gradually tapering into short style, style c. 0.5 mm long, stigma capitate. Fruit blue-black, narrowly ovoid-ellipsoid (capsule halves incurved after dehiscence giving the capsule a subspherical appearance), 5-6 by 2 mm, including c. 0.5 mm long persistent style, calyx persistent (star-like). Seeds irregular in shape, narrowly ellipsoid, laterally flattened, 1.5-2 by 0.5 mm.  Pollen grains isopolar monads, oblate spheroidal. Amb subcircular. P = 22.7 µm, E = 23.3 µm, P/E = 0.97. Aperture system 3-colporate. Ectoapertures long colpi with ± acute ends. Endoapertures ± circular, costate pori of 3-4 µm, often at one side with a small triangular extension through the costa. Exine c. 1.5 µm thick. sexine thinner than nexine; infratectum distinct, columellate. ornamentation perforate-reticulate to indistinctly rugulate; lumina irregular in shape, up to 1.5 µm, relatively small along the colpi; muri up to 1.5 µm wide.

Methods
Acetolysed pollen studied with LM and sEM.

Comment
The pollen of Hoffman & Henkel 3222 fits in Curtia as well as in Tapeinostemon. Both genera have rather diverse pollen, especially with regard to the exine surface. Curtia pollen is perforate/scabrate or perforate-reticulate to rugulate, while Tapei nostemon pollen is perforate to reticulate, sometimes indistinctly rugulate (Maguire, 1981, and own observations). Remarkably, the pollen of Tapeinostemon spenneroides shows perforate and reticulate ornamentation (Maguire, 1981) as well as rugulate ornamentation (own observation). Therefore, pollen morphology should be included in a comprehensive treatment of both genera (see also Maguire, 1981: 351).

ACknoWLEdGEMEnTs
The authors thank L.Y.Th. Westra for his help with the Latin description, P.J.M. Maas for fruitful discussions, and B.J. van Heuven for the preparation and sEM of the pollen. They are also grateful to H.R. Rypkema for the botanical illustration. L. Cobb gratefully acknowledges financial support from the Van Eeden-fonds.