Abstract: Many restoration guidelines strongly recommend the use of local sources of seed in native plant revegetation projects. These recommendations are based on assumptions that the species used for revegetation are cross-pollinated and woody, and were originally developed for overstorey vegetation in the northern hemisphere. We challenge their validity with respect to replacing or enhancing the native Australian grass component of degraded rangelands. Firstly, obligatory cross pollination has not been recorded in any Australian native grasses except for a few dioecious species. Indeed the majority of Australian native grasses so far studied have revealed complicated breeding systems that provide flexibility allowing reproduction and genotypes to be matched to the variable Australian environment. Secondly, we argue that the genetic dissimilarity among populations of a species is not proportional to the distance between them but is more related to the environmental stresses that have been placed on those populations in the past. We therefore conclude that there is little justification for the recommendation that only local sources of seed of Australian native grasses should be used particularly for large-scale revegetation programs. We provide some general guidelines for deciding on the seed sources to use depending on the purpose of the revegetation and characteristics of the species of choice.