The National Wildlife Refuge System has an active habitat restoration program and annually seeds thousands of hectares with native plant species. The noxious weed, Canada thistle (Cirsium arvense), plagues these restorations. This study evaluates planting methodology and seed mixes with the goal of recommending optimal methods to reduce infestation of noxious weeds, especially Canada thistle, in new restorations. Three planting methods (dormant season broadcast, growing season [summer] broadcast, and growing season [summer] drill) were fully crossed with three levels of seed diversity (10, 20, and 34 species [plus a fourth level, 58 species, on the three sites in Iowa]) in a completely randomized design replicated on nine sites in Minnesota and Iowa. The propagule bank of Canada thistle was evaluated at each site. Planting occurred in winter 2004 and spring-summer 2005. Here I report on results through summer 2007. None of the planting methods or seed mix diversities consistently resulted in reduced abundance of Canada thistle. Soil texture had the strongest influence; sites with greater proportions of clay had greater frequency and cover of Canada thistle than did sandy sites. At the Minnesota study sites, the dormant broadcast planting method combined with the highest seed diversity resulted in both the greatest cover of planted species as well as the greatest richness of planted species. At the Iowa sites, planted species richness was slightly greater in the summer drill plots, but cover of planted species was greatest in the dormant broadcast plots. Richness of planted species at the Iowa sites was maximized in the high diversity plots, with the extra-high diversity seed mix resulting in significantly lower species richness. Individual species responded to planting methods idiosyncratically, which suggests that particular species could be favored by tailoring planting methods to that species.