Classification of Old World Velloziaceae

Abstract

Summary. The Velloziaceae of the Old World have been revised, a new classification, incorporating anatomical evidence, is proposed, and a key to all the genera of the family is given. This is followed by identification keys, synonymy and distribution of the members of the two Old World genera, the monotypic South African genus Talbotia, and the genus Xerophyta which contains three sections, defined on anatomical features, and 28 species, 4 of which are described as new. In preparing a treatment of the Velloziaceae for the Flora of Tropical East Africa we have found it a necessary preliminary to review not only the Old World species but also the highly controversial generic limits of the whole family. The difficulties with generic limitation have been twofold. The first involves the choice and emphasis of limiting characters. The second stems from the attempt to classify Old World species under New World genera and vice versa. The first significant generic classification was that of Seubert in 'Flora Brasiliensis' (1847) where he keyed Barbacenia as having flattened bifid filaments and Vellozia filamentous ones. In the descriptions he further noted the dorsifixed anthers and capitate or clavate stigmas of Barbacenia as against the basifixed anthers and usually horizontal 3-lobed stigma of Vellozia. All of these characters correlated with each other and with the later noted features of leaf-anatomy. Most subsequent authors have misapplied these characters or used characters that do not correlate, such as the number of stamens or the length of the perianth-tube. In 1875 Baker revised the African species of Xerophyta on the basis of its six stamens and lack ofa perianth-tube, but included similar speciesof Vellozia without noting the distinctions of filament and stigma. Later (1889, 1896) he reversed the synonymy in favour of the earlier Vellozia and this time included all species regardless of the number of stamens. Pax (1887) distinguished the genera by the number and position of the stamens. He placed all species with six stamens under Barbacenia and species with more than six under Vellozia. Baillon (1895) united all species under Barbacenia while Durand & Schinz treated all African members, including Talbotia, as Xerophyta. The first serious attempt to revise all Old World species of the Velloziaceae was by Greves (1921) who classified them all as Vellozia with characterization based on the lack of a perianth-tube. Smith (1962) also overlooked the great distinction in stigmas. Perrier (1946) pointed out, when preparing for the revision of the Velloziaceae of Madagascar, that botanists are not at all in agreement on the taxonomy of the family. However, he could recognize three genera: Xerophyta, Vellozia, and Barbacenia. Hutchinson (1934) was indeed right in suggesting that 'The family needs a careful monograph, and possibly more than two genera should be recog