Formerly, non-local seed sourcing of naturally occurring herbaceous species was the main strategy for ecological restoration in Germany. Prices for non-local seeds were significantly lower and large quantities of local seeds were unavailable. We here tested whether the genetic composition of non-local seeds (R) of Pimpinella saxifraga L. (Apiaceae) of former restoration projects can still be detected several years after application and whether it differs from the species' regional gene pool (I). A comparative population genetic analysis (AFLP) was conducted. In a common garden experiment phenological and morphological differences were investigated. We further simulated mowing to investigate, if treatments affect indigenous (I) and restored (R) populations differently. The population genetic analysis resulted in two large and quite distinct and diverse cluster that separated I- and R-site individuals, with some signatures of introgression from R to I. Three generative fitness parameters featured significant differences between individuals of the R- and I-sites. The genotypic persistence reduces the availability of niches for the local genotype and may eventually lead to genotypic competition or introgression. We therefore recommend to retain from the introduction of the species' non-local genotypes in the region to avoid genotypic competition and recommend to use the genetically diverse local genotypes for restoration purposes.Data collection time: 2011.09.01 until 2013.11.30In a comparative approach, we tested whether the genetic makeup of the former non-local seed source use of Pimpinella saxifraga L. (Apiaceae) in restoration projects at sites in Central Germany several years after application (restoration between 1994 and 2004) differs from the species' regional gene pool. In a comparative approach, we compared genetic diversities and differentiations by using amplified fragment length polymorphisms (AFLPs). Phenological and morphological differences were analyzed in a common garden experiment in which we cultivated plants from all sites. For that we sampled seed and plant material in autumn 2011 of seven populations per I- and R-site.Of each individual of each study site 60 seeds were sown in a common garden experiment at the Universities nursery in Giessen (Germany) in spring 2012.The data collections of generative and vegetative morphological variation were measured two times per week in 2013 throughout the season.