The rapid growth rate of human population, along with the public health crisis encountered in many regions, particularly in developing world, creates an urgent need for the discovery of alternative drugs. Because medicinal plants are not distributed randomly across lineages, it has been suggested that phylogeny along with traditional knowledge of plant uses can guide the identification of new medicinally useful plants. In this study, we combined different statistical approaches to test for phylogenetic signal in 33 categories of plant uses in South Africa. Depending on the null models considered, we found evidence for signal in up to 45% of plant use categories, indicating the need for multiple tests combination to maximize the chance of discovering new medicinal plants when applying a phylogenetic comparative approach. Furthermore, although there was no signal in the diversity of medicinal uses-that is, total number of medicinal uses recorded for each plant-our results indicate that taxa that are evolutionarily closely related have significantly more uses than those that are evolutionarily isolated. Our study therefore provides additional support to the body of the literature that advocates for the inclusion of phylogeny in bioscreening medicinal flora for the discovery of alternative medicines.