Plant competition may mediate the impacts of herbivory on invasive plant species through effects on plant growth and defense. This may predictably depend on whether herbivory occurs above or below ground and on relative plant competitive ability. We simulated the potential impact of above- or belowground damage by biocontrol agents on the growth of a woody invader (Chinese tallow tree,ﾠTriadica sebifera) through artificial herbivory, with or without competition with a native grass, little bluestem (Schizachyrium scoparium). We measured two defense responses ofﾠTriadicaﾠthrough quantifying constitutive and induced extrafloral nectar production and tolerance of above- and belowground damage (root and shoot biomass regrowth). We examined genetic variation in plant growth and defense across native (China) and invasive (United States)ﾠTriadicaﾠpopulations. Without competition, aboveground damage had a greater impact than belowground damage onﾠTriadicaﾠperformance, whereas with competition and above- and belowground damage impactedﾠTriadicaﾠsimilarly. Whole plant tolerance to damage below ground was negatively associated with tolerance to grass competitors indicating tradeoffs in the ability to tolerate herbivory vs. compete. Competition reduced investment in defensive extrafloral nectar (EFN) production. Aboveground damage inhibited rather than induced EFN production while belowground plant damage did not impact aboveground nectar production. We found some support for the evolution of increased competitive ability hypothesis for invasive plants as United States plants were larger than native China plants and were more plastic in their response to biotic stressors than China plants (they altered their root to shoot ratios dependent on herbivory and competition treatments). Our results indicate that habitat type and the presence of competitors may be a larger determinant of herbivory impact than feeding mode and suggest that integrated pest management strategies including competitive dynamics of recipient communities should be incorporated into biological control agent evaluation at earlier stages.