AbstractJapanese knotweed and congeners are invasive to North America and Europe and spread aggressively along rivers establishing dense monotypic stands, thereby reducing native riparian plant diversity, structure, and function. Noxious weed control programs attempt to eradicate the knotweed with repeated herbicide applications under the assumption that the system will recover to a native assemblage which will inhibit future invasions. However, eradication efficacy studies typically only measure the amount of knotweed reduced, not the reestablished species diversity or plant origins. For a community scale efficacy study, we measured vascular plant species diversity and cover in riparian areas along five rivers in Washington State, 3 to 6 years after Bohemian knotweed was initially treated with herbicide. Plant species composition was compared between riparian sites treated to remove knotweed and reference sites where knotweed was absent. Sites where knotweed had been removed had significantly more exotic species and vegetative cover than reference sites; however, native species richness and cover were greater in reference sites and areas with more overstory vegetation. The native plants observed were primarily shade tolerant and perennial, as opposed to many of the exotics, which were shade-intolerant annuals. In general, reestablishment of native and exotic vegetation was not related to pretreatment knotweed stem count, size of the invaded area, or timing of herbicide application. However, residual native tree cover was negatively correlated with initial knotweed stem count. Monitoring the success of restoration objectives (e.g., native plant reestablishment or increased species diversity) and characterizing associated habitat features following knotweed eradication will help in the development of site-specific protocols for successful plant community scale restoration.