Notes on Brillantaisia pubescens (Acanthaceae) with Particular Reference to Discontinuity in Flower Size

Abstract

Summary. Two varieties are recognized in Brillantaisia pubescens, one having flowers approximately three times as big as the other. No correlation can be found with other characters, nor any geographical separation. The larger-flowered variety is described as var. riparia. Some discussion is given of possible causes of the floral dimorphism. The recent transfer of B. pubescens to Hygrophila is considered unjustified after study of pollen characters. Brillantaisia pubescens T. Anders. ex Oliv. is a rather distinctive member of the genus, easily recognized by its long spreading hairs on vegetative parts and its diffuse inflorescence with broadly obovate to almost orbicular bracts. It occurs from southeastern Tanzania to Zimbabwe, and in Madagascar. The species was described in 1875 by Oliver, based largely or wholly on a specimen collected by Speke & Grant in southern Tanzania. Since Oliver noted also 'The same species occurs in Zambesi-land', we designate the Speke & Grant specimen as lectotype rather than regard it as the holotype. No dimensions were given for the floral parts, but the description was accompanied by an illustration by W. H. Fitch. Judging from the specimen still in the Kew herbarium, Fitch seems to have used some artistic licence, adding leaves which are not on the specimen, and depicting very many more flowers than are actually present and showing them considerably larger than they are. Indeed, it might even be doubted whether the illustration was actually based on Speke & Grant's collection were it not for the fact that the original sketches for the illustration were made on the same sheet as the specimen and can still be clearly seen there. This illustration has apparently misled later authors, who have assumed that the original B. pubescens had considerably larger flowers than it actually has. Vatke in 1885 described a closely allied species, B. rutenbergiana, from Madagascar, differing only in the flowers being three times as small. In 1895 Lindau described a further species, B. anomala, from Mozambique, again distinguishing it from all other species by its smaller flowers. In the Flora of Tropical Africa, Burkill (1900) reduced B. anomala to synonymy of B. pubescens, clearly recognizing that his concept of the species included plants with large and small flowers but making no formal taxonomic separation between them. Later Benoist (1939), in his first account of the Acanthaceae of Madagascar, accepted Burkill's broad view of the species but recognized the small-flowered Madagascan material as a variety, var. rutenbergiana, based on Vatke's earlier species. We have re-examined the material in the Kew Herbarium, and find that on the basis of flower size it can be divided into two quite discrete groups with virtually no intermediates. We are unable, however, to find any other