The possibility of investigating the petrology and geochemistry of the upper mantle is discussed, using evidence from accidental xenoliths entrained in basaltic and kimberlitic magmas. Geochemical studies of metasomatized xenoliths enable likely metasomatic fluids to be characterized; following this the origin of these fluids is discussed. Most kimberlites intrude cratonic regions, whereas basaltic rocks occur in most continental tectonic regimes. Hence, a knowledge of the nature of different lithospheric environments is necessary to predict conditions conducive to kimberlite generation; geothermobarometry can be of assistance here. In conclusion, Australian crust/mantle transitions are discussed.-J.M.H.