The greater Cape Floristic Region

Aim The Cape Floristic Region (CFR) (Cape Floristic Kingdom) is currently narrowly delimited to include only the relatively mesic Cape fold mountains and adjacent intermontane valleys and coastal plains. We evaluate the floristic support for expanding the delimitation to include the whole winter-rainfall area (arid and mesic climates) into a Greater CFR. Location Southern Africa, particularly the south-western tip. Methods The initial divisive hierarchical classification analysis TWINSPAN used the presence/absence of vascular plant genera to obtain major floristic groupings in southern Africa. For the more detailed analyses, we scored the flora as present/absent within a set of centres, among which the floristic relationships were investigated (agglomerative methods, UPGMA and minimum spanning trees). These analyses were conducted with species, genera and families separately. The centres were grouped into five regions. The species richness and endemism was calculated for the centres, regions and combination of regions. The dominant floristic components of each region were sought by calculating the percentage contribution of each family to the flora. Results The divisive method showed that the winter-rainfall areas are floristically distinct from the rest of southern Africa. The species- and generic-level analyses revealed five regions: CFR, Karoo Region, Hantam-Tanqua-Roggeveld Region, Namaqualand Region and Namib-Desert Region. The CFR has the highest endemism and richness. However, the combination of the CFR, the Hantam-Tanqua-Roggeveld Region and the Namaqualand Region results in a higher total endemism. Combined, these three regions almost match the region delimited by the TWINSPAN analysis, and together constitute the Greater CFR. Main conclusions The CFR constitutes a valid floristic region. This is evident from the endemism and the distinctive composition of the flora. However, the total endemism is higher for the whole winter-rainfall area, and this supports the recognition of the larger unit. If floristic regions are to be delimited only on endemism, then the Greater CFR is to be preferred. If floristic regions are delimited on the composition of their floras at family level, then the support for such a grouping is weaker.