Abstract
Aim Four factors (time, dispersal, extinction and speciation) shape the differences in species richness within a clade between two regions. The southern African summer‐rainfall ‘Highveld’ is less species rich than the adjacent winter‐rainfall ‘Greater Cape Floristic Region’. In many clades, this may be the result of a Cape origin of these temperate clades, with later dispersal to the cool uplands of the Highveld. We explored an alternative scenario of a widespread southern African Miocene temperate flora, which survived in the Cape. We used these four factors to seek the more likely scenario for the Danthonioideae. Location Southern Africa. Methods We used a rate‐corrected and dated phylogeny including 118 of 134 species of African Danthonioideae (Poaceae), and assigned the species to be either winter rainfall (Cape) or summer rainfall (Highveld). We inferred speciation and extinction rates using TreePar and bamm for the whole clade and also for each rainfall regime, transition direction using BayesTraits and Lagrange , and integrated speciation–extinction–migration using Bi SSE . Results We found no differences in the speciation rates between the two areas, and no clear indication of the ancestral area of the clade. However, one of the most likely Bi SSE models found a lower extinction rate in the Cape, and another found a higher transition rate from summer rainfall to winter rainfall. Main conclusions We found more support for the scenario of a southern African rather than a Cape origin of the danthonioids. The winter rainfall Cape became differentiated from the rest of southern Africa in the late Miocene. Late Miocene extinction, possibly driven by the expansion of C 4 grasses, may have decimated the Highveld lineages, whereas the Cape lineages continued to diversify. Exploring some other Cape clades may reveal more instances of the Cape acting as a ‘museum’ of southern African temperate diversity.