The Genus Eugenia L. (Myrtaceae) in East Africa

Abstract

Summary. The genus Eugenia is revised for East Africa. Four new species, E. toxanatolica, E. mujindiensis, E. tanaensis, and E. thikaensis, and one new subspecies, E. capernsis subsp. multiflora, are described. Preparation of an account of the family Myrtaceae for the Flora of Tropical East Africa is made difficult not only by the presence in the area of some two hundred introduced species of the family (mainly from Australia), many so widely grown or even naturalized that they must be included, but also by the intrinsic complexity of the two indigenous genera Eugenia and the ecologically very important Syzygium Gaertn. No revision of the African species of Eugenia as a whole has been carried out although Engler (1921) gave a brief conspectus in key form of the 50 species then recognized. G. J. H. Amshoff wrote accounts for Angola (1970), French Equatorial Africa (1958) and Gabon (1966). I have found the annotations of material she examined during work on these accounts often very helpful. F. White (1977, 1978) after an admittedly lengthy study of the southern African species considered many taxa previously always treated as valid species no more than subspecies of Eugenia capensis (Eckl. & Zeyh.) Harv. ex Sond.; his approach was, despite examination of a great deal of material, highly theoretical and preconceived ideas were not dislodged by the abundance of available specimens. In this genus specific differences are indeed slight but correlated with habit and habitats. Neither Van Wyk (1979) nor Arnold & De Wet (1993) mention White's work and it is clear that it is not accepted in South Africa. A study of the South African taxa confirms that this is justified. In only one case is White's treatment relevant to East Africa and that is in relation to E. capensis being widespread up the east coast of Africa; even here it is disturbingly variable. The situation is rather worse in the Flora Zambesiaca area. At least in South Africa there had been previous revisions and accounts of the genus to fall back on but further north little had been done. Once again White has considered very disparate material over a wide area to be subspecies of E. capensis; e.g., the name E. nyassensis Engl., based on an unlocalised specimen at Kew from Malawi, has been used for many specimens not one of which shows the characteristic calyx and stem indumentum. None of the inland populations appear to me sufficiently related to the coastal ones to allow them to be treated as subspecies. It has seemed to me