Biological invasions are associated with declining biodiversity in many ecosystems, but it is often unclear if invasions are directly responsible for such changes, or if invasions are a symptom of environmental degradation. In addition, the mechanism underlying the effects of many invasive species is unknown. To determine if invaders are driving changes in invaded communities, and to identify causal mechanisms, studies that manipulate the presence of invaders are needed. We experimentally introduced the invasive grass Microstegium vimineum into replicated field plots that had been planted with the native woodland herb Senna hebecarpa. After 3 y, we quantified Senna establishment, growth and reproduction. We then conducted a greenhouse experiment to determine if changes in Senna success were due to alteration of soil microbial communities or nutrient depletion in invaded plots. Microstegium-invaded plots had 74% fewer Senna plants and Senna growing in invaded plots were 21% shorter and weighed 64% less than in control plots. The proportion of Senna plants that reproduced was 67% lower and plants produced 78% fewer seeds on average in invaded than in control plots. In contrast to the field results, there were no differences in the growth of Senna when grown in Microstegium-invaded or control soil in the greenhouse, and the invasion treatment did not alter the effects of soil sterilization or fertilization. Further, we found no evidence that Microstegium success is determined by feedbacks with the soil community. Our results demonstrate that Microstegium has negative effects on a native species, but we found no evidence that the suppressive effects of Microstegium invasions are mediated by plant-soil interactions in invaded areas.