Rare species, Restionaceae, and the Cape flora

Aim I test whether species range size variation is driven by the age of the species, their intrinsic traits, environment or by historical‐geographical patterns (e.g. the location of climatic or environmental refugia). Location Greater Cape Floristic Region (GCFR) of South Africa. Taxon Angiosperms, Restionaceae (restios). Methods I determined the area of occupancy and extent of occurrence of all restios in the GCFR, and derive a list of “single‐site endemics” (SSE). I inferred a rate‐corrected and dated phylogeny which includes 98% of the species. I explored the correlates of range size variation with phylogenetically corrected regressions, focussing on likely candidate traits (dispersal and fire survival modes), habitats (soil drainage, elevation), climatic variables (rainfall and temperature) and history (age of species). I calculated the species richness, community mean range sizes and numbers of SSE for 49 sites and 938 10 m diameter circular relevés across the GCFR, and sought correlates of these, specifically testing for historical‐geographical patterns in relationship to the distance from the centre of the region. Results Narrow‐range restios are correlated with higher elevations, rainfall and rainfall seasonality; are found more likely in marshes and seepages; and are more likely killed by fire and have to re‐establish from seed after fire. Wide‐ranging species are found in areas with lower rainfall, at lower elevations, in well‐drained habitats, and resprout after fire. Community mean ranges sizes are larger towards the eastern part of the GCFR; and relevé data show that sandstones, shale, laterite and granites are associated with wider distribution ranges, as are soils with pebbles or sandy soils without rocks, and limestones with smaller community mean ranges than other bedrock types. The number of SSE is largely correlated with species richness, except that SSE are more likely at higher elevations. Main Conclusion Range size variation in the restios is largely driven by history (expressed as the distance from the Kogelberg region), suggesting that the Kogelberg region may have been most climatically stable during the Neogene and Pleistocene. The second‐most important driver is the environment (i.e. rock type, elevation, rainfall), generally with spatially restricted habitats hosting range‐restricted species.