Abstract. Soils in the south-west of Western Australia and South Africa are among the most phosphorusimpoverished in the world, and at the same time both of these regions are Global Biodiversity Hotspots. This unique combination offers an excellent opportunity to study root adaptations that are significant in phosphorus (P) acquisition. A large proportion of species from these P-poor environments cannot produce an association with mycorrhizal fungi, but, instead, produce 'root clusters'. In Western Australia, rootcluster-bearing Proteaceae occur on the most P-impoverished soils, whereas the mycorrhizal Myrtaceae tend to inhabit the less P-impoverished soils in this region. Root clusters are an adaptation both in structure and in functioning; characterized by high densities of short lateral roots that release large amounts of exudates, in particular carboxylates (anions of di-and tri-carboxylic acids). The functioning of root clusters in Proteaceae ('proteoid' roots) and Fabaceae ('cluster' roots) has received considerable attention, but that of 'dauciform' root clusters developed by species in Cyperaceae has barely been explored. Research on the physiology of 'capillaroid' root clusters formed by species in Restionaceae has yet to be published. Root-cluster initiation and growth in species of the Cyperaceae, Fabaceae and Proteaceae are systemically stimulated when plants are grown at a very low P supply, and are suppressed as leaf P concentrations increase. Root clusters in Proteaceae, Fabaceae and Cyperaceae are short-lived structures, which release large amounts of carboxylates, briefly, at a particular stage of root development. The rates of carboxylate release are considerably faster than reported for non-specialized roots of a wide range of species. Root clusters play a pivotal role in mobilization of P from P-sorbing soil. Because the world P reserves are being depleted whilst vast amounts of P are stored in fertilized soils, there is a growing need for crops with a high efficiency of P acquisition. Some Australian and African native species as well as some existing crops have traits that would be highly desirable for future crops. The possibilities of introducing P-acquisition-efficient species in new cropping and pasture systems are explored. In addition, possible strategies to introduce traits associated with a high P-acquisition efficiency into future crop species are discussed.