SYSTEMATICS OF THE SOUTHERN AFRICAN GENUS GEISSORHIZA (IRIDACEAE -IXIOIDEAE)1 PETER GOLDBLATT2 ABSTRACT Geissorhiza is a large genus of small, corm bearing perennials endemic in the Cape Province of South Africa and is restricted to the winter rainfall area of the south and west coasts and immediate interior. The genus is defined by its asymmetric corms with hard, often woody tunics, herbaceous bracts, and long style with short recurved branches. It is separated by style morphology from the closely related Hesperantha, in which the style is included in the perianth tube and the style branches are long and spreading, and, in addition, the anthers of Hesperantha are introrse at anthesis. Eighty-one species are recognized in this study, 29 described here for the first time. This represents an increase of some 25 over the 52 species of true Geissorhiza admitted by Foster (1941) in his revision of the genus and the eight species placed by Lewis (1941) in Engysiphon, a genus here reduced to synonymy. A novel subgeneric classification is proposed in which two subgenera, Geissorhiza (34 spp.) and Weihea (47 spp.) are recognized, each including several sections. Geissorhiza is seen as a late Tertiary genus that probably differentiated in upland southeastern Africa when much of southern Africa including the Cape Region was still heavily forested. The continuing deterioration ofthe climate in the Pliocene and Pleistocene that culminated in the establishment of a Mediterranean climate in the southwestern Cape probably gave the impetus to the extensive radiation in the mountains and western coastal belt in this area, where most species, including ail of the more specialized taxa, occur. Primitive taxa are concentrated along the relatively well-watered southern Cape coast, where little speciation has occurred and the primitive taxa appear to have differentiated slowly by a process of phyletic evolution. The geography of Geissorhiza is analyzed in detail and several important centers of endemism, corre-sponding closely with Weimarck's phytogeographical subdivisions are identified. The major trends in the evolution of Geissorhiza include: the development of imbricate corm tunics from the basic con-centric type; the reduction or extreme elongation of the perianth tube; the development of floral zygomorphy; and the elaboration of the leaf lamina, which may become several ribbed and grooved (section Geissorhiza), or may have raised and winged ciliate margins and a winged midrib (section Ciliata) or thickened margins and midrib (especially sections Angustifolia and Ixiopsis). Cytological studies have brought the number of species in which the chromosome number is known to 41, over half the genus. Basic chromosome number is x = 13, and while most species are diploid, polyploidy has been recorded in a few species including the triploid vegetative apomict, G. bolusii. The conser-vation status of Geissorhiza is reviewed and several seriously threatened species are identified including almost all those endemic to the coastal plain between Cape Town and Piketberg, an intensively farmed area rich in species. Geissorhiza is a large genus ofmedium to small, sorhiza is a member of Ixioideae, a predomi-corm bearing plants, restricted entirely to the nantly African subfamily, strongly developed inwinter rainfall region ofthe south and west coasts southern Africa, and is distinguished by its woodyof southern Africa. It is centered in the moun-to papery (rarely fibrous) corm tunics, asym-tains and western coastal belt ofthe southwestern metric corms, herbaceous floral bracts, and long,Cape, with species extending as far east as Gra-exserted style (included in two species) with shorthamstown in the eastern Cape and as far north recurved style branches. It is most closely relatedas Steinkopf in northern Namaqualand. Geis-to Hesperantha (ca. 55 spp.-Goldblatt, 1982a, SSupport for this research from the U.S. National Science Foundation grant DEB-78-10655 and 81-19292 isgratefully acknowledged. I also wish to thank the South African Department of Forestry and the Cape Departmentof Nature Environmental Conservation for their support of this project and for collecting permits. For theirassistance in the field or in providing me with material for study I want to thank Dee Snijman, ComptonHerbarium, Cape Town; Ion Williams, Vogelklip, Hermanus; J. H. J. Vlok, Department of Forestry, George;Neil MacGregor, Glenlyon, Nieuwoudtville; Mike Viviers, Department of Forestry, Cedarberg; and especiallyElsie Esterhuysen, Bolus Herbarium, University of Cape Town, who helped me extensively with the rare highmountain species and after whom two species have been named. The support and hospitality provided by JohnRourke and his staff, Compton Herbarium, Kirstenbosch, Cape Town in the course offield work is acknowledgedhere with gratitude. I also wish to thank Margo Branch for the many excellent illustrations made for this study. 2B. A. Krukoff Curator of African Botany, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166.ANN. MISSOURI BOT. GARD. 72: 277-447. 1985.