Notes on African Rubiaceae: Vanguerieae

Abstract

Summary. This precursor paper in connection with the account of the Rubiaceae for Flora of Tropical East Africa (Part 2) deals with the remainder of the Vanguerieae with the exception of Canthium and genera formerly confused with it which forms the subject of a separate account by D. M. Bridson. The genus Rytigynia is revised for the areas of Flora of Tropical East Africa and Flora Zambesiaca with the description of 15 new species. Apart from this, one new genus, 5 new subgenera, 4 new species, 6 new subspecies, three new varieties, 31 new combinations and one new name are published. The following notes result from continuing work on the family Rubiaceae in connexion with the second volume of this family for the Flora of Tropical East Africa and complete the Vanguerieae with the exception of the genus Canthium and its allies which are being studied by D. M. Bridson. The Vanguerieae are an extremely natural tribe and the underlying pattern of similarity common to all the species makes classification exceedingly difficult; in fact, the delimitation of genera in the tribe has been the most difficult problem encountered during writing this flora account. A detailed generic revision* utilizing modern techniques would delay the appearance of the volumes yet further and only as much has been done as is necessary to produce a reasonably practical account. Pragmatism, not logic, has dictated most of the conclusions reached in this account. I commented on the naturalness of the tribe and poorness of the characters used for generic delimitation many years ago (1958: 246-7) but no improvements have been suggested since that date. It is instructive to follow the history of the tribe in the literature. Canthium Lam. was the first genus of the tribe to be described (Lamarck 1785: 602) based on two very diverse species C. coronatum Lam. (a synonym of Catunaregam spinosa (L.f.) Tirvengadum a member of the Gardenieae) and C. parviflorum (a synonym of Canthium coromandelianum (Burm.f.) Alston) accepted as the lectotype of the genus; the second was VangueriaJuss. (1789: 206) based on material collected by Commerson but not on a described species. These two genera, one defined partly by its 2-locular ovary and the other by a 5-locular ovary, dictated the use of ovary locules in all future classifications. It is also important that although the tribe is basically an African taxon neither of these early genera was based on material collected in Africa. Dumortier (1829: 32) first mentions a tribe Vanguerieae ('Vaugnerieae') attributed to A. Richard and made to comprise Hamelia, Evosmia, Mitchella, 'Vaugneria' and Laugeria. Richard (1829) had no clear concept of the tribe either, placing Canthium in his Coffeaceae and Ancylanthus, Cuviera, Pyrostria and Vangueria in Guettardaceae, the separation being based on the bilocular ovary of the first as against the plurilocular ovary of the second group. De Candolle (1830: 341-622) follows the same scheme. Bentham (in