Romulea pilosa and R. quartzicola (Iridaceae: Crocoideae), two new species from the southern African winter rainfall region, with nomenclatural corrections

Abstract

Romulea pilosa J.C.Manning & Goldblatt and R. quartzicola J.C.Manning & Goldblatt are two narrow endemics from the southern African winter rainfall region. An early, fragmentary collection of R. pilosa from Riviersonderend lacked the diagnostic corm and was thus mistakenly associated with R. tetragona (sect. Ciliatae) as var. flavandra M.P.de Vos because of the highly distinctive pilose, H-shaped leaf. The rediscovery of the taxon in the wild shows it to be a previously unrecognized member of sect. Aggregatae, distinguished by its unusual foliage and bright orange flowers. R. quartzicola was grown to flowering from seeds collected from quartz patches in southern Namaqualand and proved to be a new species of sect. Ciliatae, distinguished by its early flowering, short, subclavate leaves with reduced sclerenchyma strands, and bright yellow flowers with short bracts. R. neglecta M.P.de Vos, a rare endemic from the Kamiesberg in Northern Cape, is a later homonym for the Mediterranean R. neglecta Jord. & Fourr., and the earliest name for this plant is shown to be R. speciosa (Ker Gawl.) Baker, typified by an illustration in Andrews’ The botanist’s repository. An epitype is designated to fix the application of the name. We have also examined the type illustration of R. pudica (Sol. ex Ker Gawl.) Baker, hitherto treated as an uncertain species, and are confident that it represents the species currently known as R. amoena Schltr. ex Bég., and takes priority over it as being the earlier name. The type of R. reflexa Eckl., a new name for the later homonym I. reflexa Thunb. and the basionym of R. rosea var. reflexa (Eckl.) Bég., has been mistakenly identified as an Ecklon collection but is in fact the collection that formed the basis of Thunberg’s I. reflexa. This collection is actually a form of R. flava Lam., and the name R. rosea var. reflexa is thus moved to the synonomy of that species. The variety currently known under this name should now be known as R. rosea var. muirii (N.E.Br.) Goldblatt & J.C.Manning. Finally, the protologue of R. parviflora Eckl., until now treated under the synonomy of R. obscura Klatt var. obscura, is in fact consistent with R. rosea var. australis (Ewart) M.P.de Vos, and we therefore include the name in the synonomy of the latter.