Search

J. Lovett, whose remarkable talent for collecting new species of plants during his surveys of forest areas in Tanzania continues to increase our knowledge of the flora, handed me for naming a flowering gardenioid shrub apparently quite unlike any previously known. Study of Keay's (1958) revision of the African plants previously referred to Randia or Gardenia indicates that its closest affinity is with a species collected twice last century in Fernando Po but apparently never refound there, namely Mitriostigma barteri Hook. f. ex Hiern which more recently has been rediscovered in Cameroun. The details of floral morphology agree well, particularly the short rounded corolla-lobes overlapping to the left, anthers just included in the widened part of the corolla-tube, pollen grains in tetrads, shortly bifid stigmatic club, narrow calyx-lobes, and the inflorescences borne at successive nodes on one side only and alternating from one side to the other..The new plant differs in having a much longer purple corolla and red calyx, Mann describing those of M. barteri as brown and green respectively. Letouzey, however, on the field-note to his collection 11168 from Cameroun describes the corolla as red outside and rose inside, so even at this level the agreement is good. Later a second specimen, fruiting, was found in the Oxyanthus covers allowing completion of the description and just as this paper was completed, Prof. A. Borhidi of the Hungarian Academy of Sciences visited Kew with a collection of Rubiaceae which contained a second fruiting gathering.

Randia naucleoides S. Moore in Journ. Bot. 44: 83 (1906). Type: Uganda, Entebbe, Bagshawe 771 (holotype BM). Bertiera capitata De Wild. in Ann. Mus. Congo V, 2: 169 (1907); Wernh. in Journ. Bot. 50: 163 (1912); Verdc. in Kew Bull. 7: 359 (1952); N. Hall6 in Not. Syst. 16: 285, fig. 3/a (1961) & in Fl. Gabon 17, 2' partie: 64, pl. 14 (1970). Type: Zaire, Equatoria Province, Eala, Laurent 131 (lectotype BR, not seen).

The species Rothmanniafischeri (K. Schum.) Bullock has, since Keay's treatment in Bull. Jard. Bot. Brux. 28: 50 (1958), generally been regarded as polymorphic and although the variation corresponds with geographical distribution no formal attempts at splitting it have hitherto been made. While reconsidering the differences between R.fischeri and the less variable and well known South African species R. capensis Thunb. it became apparent that the presence of sinuses between the calyx-lobes of R. capensis was the most reliable character for separation. Previously a few anomalous specimens of R. fischeri (e.g. Malawi, Buchanan 1147 & Zimbabwe, Eyles 7652) with well developed linear calyx-lobes typical of R. capensis has appeared to be intermediate with R. capensis Fig. 1E. Furthermore this character of well developed sinuses was also found to separate the Kenyan and Tanzanian coastal element previously included in R.fischeri. Here I regard this taxon as a distinct species and the combination Rothmannia ravae is made to accommodate it and a full description and citations are given. The remaining specimens (those lacking sinuses between the calyx-lobes or with a truncate calyx-limb) have been classified into three subspecies ofR.fischeri. Keay cited Gardenia annae E. P. Wright var. moramballa Heirn in synonymy with R.fischeri but did not mention the epithets R. fischeri var. major K. Schum. or Randia ravae Chiov.