Gyrocarpus angustifolius (Hernandiaceae), a Distinct Northeast African Tree Species

Abstract

Summary. Evidence for treating G. hababensis Chiov. var. angustifolius Verdc. as a distinct species is presented, and the new combination G. angustifolius (Verdc.) Thulin is made. Gyrocarpus hababensis Chiov. was described on the basis of three collections from Eritrea by Chiovenda (1911). Verdcourt (1967) recorded the species also from eastern Ethiopia, northern Somalia and northern Kenya, and also described var. angustifolius Verdc. from eastern Ethiopia and northern Kenya. Two collections of var. angustifolius from northern Somalia were also cited, but almost certainly both actually originate from Ethiopia. Verdcourt pointed to the oblong to elliptic-oblong leaves and the velvety fruit body as distinguishing features for the variety. He also stated that "further material of this plant is needed to indicate its true status", a particular problem being that in many cases either leaves or fruits are missing in the collections. Kubitzki (1969) in his monograph of the Hernandiaceae followed the taxonomy of Verdcourt and cited the same collections. However, he also lectotypified G. hababensis and explicitly treated Gillett 12594 from northern Kenya as a mixed collection, and added the pinnate venation of the leaves to the distinguishing features of var. angustifolius. Verdcourt (1985) in his treatment for Flora of Tropical East Africa again treated the narrow-leaved plant as var. angustifolius, now giving its distribution as northern Kenya, Ethiopia and southern Somalia. However, apparently Kubitzki's lectotypification was overlooked, and Gillett 12594 was cited as var. hababensis without comments. Verdcourt also commented on the status of var. angustifolius and reproduced the opinions ofJ. B. Gillett and M. Gilbert, who, on the basis of field experience, considered var. angustifolius to be a distinct species. Still, Verdcourt preferred to follow the treatment in Kubitzki's monograph. I have also had the opportunity to study both taxa in the field in northeast Kenya and var. angustifolius in central and southern Somalia. Furthermore, material of both taxa is now available which is considerably richer than that seen by Kubitzki, and if it is understood that Gillett 12594 from northern Kenya and Ellis 374 from eastern Ethiopia are mixed collections, the pattern becomes clear. Var. angustifolius consistently differs from var. hababensis in its oblong to narrowly ovate leaves which are up to 2 -2 cm wide, sparsely to fairly densely pubescent with short appressed hairs, and have a very prominent, ridge-like