Seed for restoration projects has traditionally been sourced locally to “preserve” the genetic integrity of the replanted site. Plants grown from locally sourced seeds are perceived to have the advantage of being adapted to local conditions, and the use of local provenance is a requirement of many restoration projects. However, the processes of climate change and habitat fragmentation, with the subsequent development of novel environments, are forcing us to reconsider this basic tenet of restoration ecology. We tested the “local provenance is best” paradigm, by comparing the performance of plants grown from local with non-local seed sources within a common garden experiment. We selected six species representing a range of growth forms (Acacia falcata, Bursaria spinosa ssp. spinosa, Eucalyptus crebra, E. tereticornis, Hardenbergia violacea and Themeda australis) from an assemblage known as the Cumberland Plain Woodland, a threatened community in western Sydney. Multiple provenances were collected from within the range of each species and grown at two field sites on the Cumberland Plain. Growing time varied between species and ranged from 7 months to 2 years. With the exception of B. spinosa, and to a lesser extent T. australis, we found little evidence that local provenance plants were superior to distant provenances in terms of survival and establishment.