The speciation process that underlies recent, rapid radiations of plants is controversial, and suggested mechanisms range from pollinator or ecological niche differentiation to allopatry and nonadaptive divergence. Phylogenetic approaches to locating the most appropriate speciation models have been constrained by the low levels of molecular divergence between recently diverged species, which are typical of recent, rapid radiations. In this issue of Molecular Ecology, Rymer et al. (2010) used coalescence analyses of sequence data and genome scans of Amplified Fragment Length Polymorphism (AFLP) loci to demonstrate that in a species complex in the irid genus Gladiolus, a member of the hyper diverse Cape flora of southern Africa, speciation is a gradual process. Older divergences are genetically more differentiated, and show a greater difference in flowering time and floral morphology, than taxa that diverged more recently. There is no evidence of any abrupt events. Gene flow is limited by shifts in flowering time and floral morphology; thus, by pre-zygotic rather than by post-zygotic mechanisms, these evolved together with the occupation of somewhat different habitats. This research gives the first critical insight into how the remarkable diversity in a diversity hotspot could have arisen. More importantly, it demonstrates that the speciation process in recent, rapid radiations is tractable and can be investigated with suitable genetic tools.