Studies in Coffea and Psilanthus (Rubiaceae subfam. Cinchonoideae) for Part 2 of 'Flora of Tropical East Africa': Rubiaceae

Abstract

During the course of the revision of the genus Coffea L. for the F.T.E.A. Rubiaceae vol. 2 it became apparent that some of the East African species dealt with by Chevalier (Les Caf6iers du Globe 3 (1947)) were in need of revision, while other epithets had subsequently been misapplied. Not only was the number of indigenous species of Coffea sensu stricto very much larger than at first anticipated, but one species of Coffea falling outside the definition of Coffea sensu stricto (see pages 852-4) and three species of the closely related genus Psilanthus (see pages 854-8) were found. Although adequate material to allow full formal description of five new species of Coffea is known, an additional 11 taxa (probably worth specific rank) are also included. Because of the considerable economic importance of Coffea such novel taxa have been treated in as much detail as possible so as to permit informed field studies which could lead to re-collection of both good herbarium specimens and living material for experimental culture. As well as fulfilling the necessary nomenclatural requirements for the description of new taxa and the reassessment of existing taxa, this paper includes dichotomous and multi-access keys and a character list which will enable it to be used as a field guide by personnel with a moderate knowledge of botany. The three commonly cultivated species C. arabica, C. canephora and C. liberica have not been dealt with fully in this paper, although they have been included in the keys and the character list. In addition to cultivated plants, wild plants can be found as follows: C. arabica: SE Sudan (Boma Plateau); SW Ethiopia & N Kenya (Mt Marsabit) (fide Friis in Proc. 9 Plenary Meeting of A.E.T.F.A.T.: 63 (1979)). C. canephora: Sudan, Uganda & Tanzania (Bukoba District). C. liberica (closest to C. liberica var. dewevrei (De Wild. & Th. Dur.) Lebrun including C. excelsa Chev.): Sudan & Uganda. Since most of the characters separating the species in Coffea are generally weak and variable and so many taxa are imperfectly known, definite opinions on the relationships between species can only be made with difficulty. Furthermore, it would be unwise to recognise infrageneric taxa in this paper as the West African and Madagascan species have not been considered. In my opinion Chevalier's subsection Mozambicoffea is artificial, especially with regard to the inclusion of C. eugenioides which is much closer to C. canephora