* One of the key questions in ecosystem restoration is the choice of seed material for restoring plant communities. More and more scientists and practitioners are currently advocating the use of regional seed sources, based on the argument that plants are often adapted to local or regional environmental conditions, and thus, regional seed sources should provide the best restoration success. However, there is still substantial debate about this approach, partly because of a lack of solid empirical data. * We conducted a multispecies transplant experiment in which we compared the performance of eight seed origins of seven plant species frequently used in grassland restoration in four common gardens across Germany. * We found that, on average, plants of regional origins produced 10% more inflorescences and 7% more biomass than those of foreign origins. There were substantial differences among species in the strength of these effects, but in the majority of the study species fitness decreased with increasing geographical distance of seed origins or with increasing climatic differences between plant origins and experimental sites. * In addition to these effects on plant fitness, increasing geographical or climatic distances of origin were often also correlated with increasing differences in plant phenology. Since phenology is important for biotic interactions, especially with pollinators and seed predators, using foreign seed sources may have cascading effects on local ecosystems. * Synthesis and applications. Genetic differentiation is widespread in grassland species and often shows the patterns of regional adaptation. Our study thus supports the use of regional seed sources in restoration. Moreover, using non-regional seed sources in grassland restoration may not only decrease the performance of plants, but it will likely also affect their biotic interactions.