A New Genus Haplocoelopsis (Sapindaceae) from East and Central Africa

Abstract

The distinctive tree described below, appears on present evidence to have been first collected in October 1937 by Ivan Dale, then a forester in the Coastal Province of Kenya. Further material was obtained in the Pugu Hills (Vaughan in 1940) and Newala (Gillman in 1943). The plant was determined as a new species of Haplocoelum at Amani by Greenway and was included as such by Brenan in his checklist (1949: 558). It was recognised at Kew that it was unlikely to belong to Haplocoelum since petals were present and several sheets were loaned to P. W. Leenhouts who annotated them as "prob. Cupanieae, nov. gen. aff. Eriocoelum". My work has confirmed the position of this new genus and it is described below. Of the African Sapindaceae with stipule-like lower leaflets (pseudo-stipules), Blighiopsis Van der Veken and Lepisanthes Blume can be immediately excluded since the gynoecium in both has a basal style and the fruit consists of individual mericarps. Haplocoelopsis has a terminal style and presumably a two-valved capsule. The tribe Cupanieae contains the genus Eriocoelum Hook. f. with pseudo-stipules and also Aporrhiza Radlk. with a two-locular ovary. Some members of the tribe have the sepals valvate or just imbricate in bud but others have sepals opening precociously as in the adjacent tribe Schleichereae. Pollen characters are useful in the family (Muller & Leenhouts 1976) but unfortunately adequate male flowers are apparently not yet known for Haplocoelopsis. The distribution of the sexes is very varied in the Sapindaceae and difficult to make out in many taxa but the presence of pedicel remains on a female-flowered thyrse and comparison with other members of the Cupanieae would suggest that Haplocoelopsis is monoecious and progresses from a male to a female stage within the same inflorescence.