Questions:Restoring native-dominated plant communities often requires controlling invasive species, reintroducing native species, and implementing continued management practices. Can single herbicide applications to control Centaurea stoebe L. encourage establishment of seeded native species more effectively than a single mowing? Can annual hand pulling to control C. stoebe favor the persistence of seeded native species?Can mid-spring burning reduce C. stoebe and increase native forbs and grasses? After eight years, will the restored plant communities differ from those in untreated areas?Location: Bass River Recreation Area, Ottawa County, MI, USA. Methods:We studied the effects of site preparation (mowing, clopyralid, glyphosate), hand pulling of C. stoebe, and burning on restoring native plant communities on a C. stoebe-infested site. Over eight years, we quantified the development of the plant communities on plots seeded with native grasses and forbs, and report on the second four years here.Results: Native-dominated plant communities developed using both herbicides, but while clopyralid provided longer control of C. stoebe, clopyralid-treated plots had fewer native species than glyphosate-treated plots. Native-dominated plant communities also developed on plots that were only mowed once before seeding, achieving similar native species richness as the glyphosate treatment. Hand pulling controlled C. stoebe, burning increased relative cover of native graminoids and decreased that of non-native grasses, and hand pulling and burning in combination increased relative cover of native forbs. After eight years, the restored plant communities had greater native species cover and richness and higher mean Coefficient of Conservatism, Floristic Quality Index, and Shannon's Diversity Index values than untreated areas. Conclusions: Site preparation, seeding, hand pulling of C. stoebe, and annual burning facilitated restoration of native-dominated plant communities on a C. stoebe-infested site. Effects accumulated over a period of eight years, illustrating the importance of continued management and monitoring as part of similar restoration efforts. Additional supporting information may be found online in the Supporting Information section at the end of the article. Appendix S1. Aerial view of the study site, showing the four experimental blocks of 12 5-m × 5-m plots each. Appendix S2. Frequency of occurrence and mean relative percent cover of common and seeded species. Appendix S3. Plant cover, diversity, and floristic quality data. Appendix S4. Knapweed lifestage density data.