Background/Question/Methods
Local extinction of native species can occur following the establishment of invasive species. These localized extinctions are often attributed to the superior competitive ability of the invading species, although experimental proof is often lacking. Moreover, there are few studies that have examined how environmental conditions (e.g., microhabitat, soil conditions, source of invasion) influence competitive exclusion by invaders. Within the prairie pothole region of the United States and Canada, the grass, Bromus inermis (smooth brome) has become established by invading disturbed prairies, and through repeated introductions for soil retention and animal graze. Our previous research has shown that local patch extinction of native prairie cordgrass (Spartina pectinata) is 8 times more likely to occur in areas heavily infested with smooth brome than areas not infested by brome. Although suggestive, these data do not constitute proof that this invasive grass promotes local extinction of cordgrass. In this study, we experimentally test the hypothesis that local cordgrass extinction is driven by competitive exclusion. Specifically, we test whether competition occurs between brome and cordgrass and whether the strength of competition varies across dominant habitat types present in the prairie landscape (mono-dominant stands of cordgrass, mono-dominant stands of brome and a mixture of native grass species). Two experiments were conducted: (1) planting of brome and cordgrass seeds into cleared plots of each habitat type to assess competitive interactions, and (2) planting of brome seeds into natural cordgrass habitats to assess colonization success of brome seed. For both experiments, plots were monitored for three years.  
Results/Conclusions
In experiment (1), brome was a dominant competitor across all habitats (F= 58.116, df =1 and p<0.001). Across all three habitats, the outcome of competition was unaffected by soil nitrogen or elevation. However, within brome habitats, the decrease in cordgrass density was greatest when soils were driest (F=4.055, df= 1 and p<0.053). Effects of habitat and edaphic conditions on brome and cordgrass seed production is currently under investigation. In experiment (2), brome seed had high germination rates in natural; cordgrass habitats, but no seedlings survived past two years or flowered. Overall, our data suggest that dispersal of brome seeds into established cordgrass habitat is likely to be met with failure. However, once brome becomes established it can competitively exclude cordgrass. We conclude that local extinction of cordgrass patches is brought about by the spread of competitively superior smooth brome.