Six new Pavetta (Rubiaceae), including three ‘litter-bin’ species from the evergreen forests of Western Africa

Abstract

Terms, definitions and measurements for leaf domatia and inflorescence parts follow Robbrecht (1988) and Manning (1996), but some clarification of definitions may be useful. Inflorescences of most Pavetta and all the following species are compound dichasia, usually with all flowers or fruits developing at about the same time. The dichasia are usually terminal on flowering branchlets of a length (measured to primary inflorescence node, defined below) and with a leaf arrangement that are, within some species groups, constant enough to be useful for identification. In P. sapoensis the flowering branchlets are highly contracted so the dichasia appear at first glance to be axillary on peduncles arising directly from leaf axils. The ‘primary inflorescence node’ is that at which the pair of largest branches of the dichasium meet the main axis, and the radius of the inflorescence (or infructescence) is measured from this primary node to the tip of the pedicel of the most distant flower or fruit, so sizes can be compared even if fruits or flowers have fallen. Although there is some limited expansion of the inflorescence axes after the flower buds are fully formed, the radius of an inflorescence with open flowers varies relatively little within species, and upper limits can be defined that vary considerably between species. The diameter of the inflorescence is approximated here as twice the maximum radius even though the inflorescences are not always strictly hemispherical. The following species have a wide range of form of bracts and bracteoles, typical for Pavetta, including: stipule-like sheathing bracts at the primary and some lesser nodes; minute bracteoles on pedicels; and ‘other bracts’ between these extremes. In some species, one or more foliar bracts (green and leaf-like, but usually smaller than leaves) may be seen at nodes in the dichasium, but even in species where they have been recorded, these foliar bracts are absent from many specimens. Pavetta flowers are characteristically 4-merous, unlike the related genus Tarenna, and there are no exceptions amongst our new species. The swollen section of the club-shaped style is called a pollen-presenter. In the ixoroid pollination mechanism (or secondary pollen presentation mechanism) typical of the tribe, the anthers release the pollen before the stigma is receptive. This release of pollen occurs in the closed bud, with pollen being deposited on the upper part of the style and/or the outside of the tightly appressed stigmatic lobes. At anthesis, the flower opens and the pollen presenter becomes exposed. The flower is then in a functionally male stage, during which pollen is available for pollinators. In the next stage (functionally female stage), the stigmatic lobes separate, exposing the papillate inner surfaces, which are receptive for pollen from other plants. The stipules have a collar section running between the petiole bases and often have an awn-like or stouter, linear apical part called the stipular awn. The stipular awn often runs into a prominent midrib-like ridge on the lower part of the stipule, but awn length excludes this and includes only the free section of awn. All long, hair-like structures forming an often dense indumentum on the inner face of the stipules of some species below may strictly be, at least in part, filiform colleters as in Ixora (de Block 1998), but are all referred to as hairs below, as their anatomy and function has not been investigated and they appear, under a dissecting microscope, to intergrade with trichomes elsewhere on the plants. Colleters are often reported to be dark in colour, whereas in the new species below the ‘hairs’ under the stipules are very pale. Corky scabs typically soon develop on the stipules of many species of subgenus Baconia Bremek., including the first three species below. Other Pavetta species, including our last three, have more leathery and triangular stipules often with a corky margin. Apparently our last three species also belong in subgenus Baconia, based on flower characters of two of them and other details. Subgenus Baconia is characterised by flowers with a bearded corolla throat and the styles that are not as long-exserted as in other subgenera (Bremekamp 1934). The five of our species for which open, mature flowers are known share these diagnostic features and the one species for which the fully mature flower is unknown has other strong affinities with species of subgenus Baconia. Bridson (1978), Bridson & Verdcourt (1988) and Manning (1996) agree that the series established by Bremekamp (1934) are of no taxonomic value in most cases. I follow these authors here by using Bremekamp’s (1934) classification for pragmatic rather than systematic reasons. A generic revision and/or further systematic investigation is likely to establish more meaningful infra-generic alliances. The notes on taxonomic affinity given below are not confined to species of any one subgenus.