The ecological life history of Microstegium vimineum, an invasive grass, was compared among four populations throughout southern Illinois. A detailed investigation of seed production was conducted at one of these populations in a secondary oak-hickory forest in Dixon Springs State Park, Illinois. In this park, M. vimineum occurs as a central, main population with several disjunct subpopulations. Size of flowering and non-flowering individuals was determined as the plants set seed. Seed dispersal was assessed using seed traps. A severe late summer drought impacted the Dixon Springs population in 1999 and few plants flowered. Plants that did flower were significantly larger than those not flowering and were limited to one small area at the edge of the core population. At Dixon Springs, the germinable seed bank was estimated from soil samples collected in Fall 1999 before flowering, and in the subsequent spring. Seedlings emerged from soils placed in a greenhouse from samples collected in the spring. No seedlings emerged from samples collected in the fall. Germinable seed had a 10 times higher density in samples collected near the soil surface than from samples collected at 5-10 cm depth. Seed dispersal occurred within the main population only. Despite the lack of seed dispersal in 1999, seedlings emerged from six of seven subpopulations in spring 2000, with the highest densities where plants had flowered in 1999. These data indicate the presence of a persistent soil seed bank for M. vimineum. Even following summer drought and discounting spring seedling emergence, the size of the seed bank increased by similar to25% in 1999. Reestablishment of the Dixon Springs population following the 1999 drought was vigorous, but the location of the largest, most fecund plants in 2000 was not the same as in 1999. In 2000, end of season survivorship of M. vimineum among all populations was 40-50%, with 90% of surviving individuals flowering. Performance of M. vimineum varied significantly within and among populations and was related to soil texture and chemistry, and overhead canopy cover. The findings of our study reflect the invasive nature of this plant. A sufficiently large seed bank and a highly plastic morphological response to local microhabitat conditions likely ensures its persistence in invaded sites.