Microstegium vimineum, an Asian annual C4 grass that is very shadetolerant, has invaded floodplains, streambanks and adjacent mesic slopes in the North Carolina Piedmont during the past 30 years. A 3-year study of its invasive characteristics revealed that M. vimineum is slow to invade undisturbed vegetation, but rapidly fills disturbed, mesic, shaded areas, such as streamsides where floods scour existing vegetation or sewer-line rights-of-way which are mown once a year. Its seeds remain viable for at least 3 years in the soil seedbank and rapidly germinate to produce a new cohort if a disturbance removes an existing cohort. On fertile floodplain sites, soil fertilization in March had no effect on seed production in October. When M. vimineum seeds were sown into existing vegetation, seed production was negatively correlated with soil potassium, calcium, silt and pH, probably because the more fertile sites also supported a denser ground vegetation layer. These qualities, in addition to its cleistogamous or apomictic reproduction, help explain how M. vimineum has spread throughout the eastern U.S. since its introduction approximately 70 years ago. INTRODUCTION Microstegium vimineum (Trin.) A. Camus is an annual grass species from Asia which recently invaded floodplains, streamsides and adjacent mesic slopes in the North Carolina Piedmont. It appears to be advancing into the existing ground cover of floodplains, which consists mainly of Japanese honeysuckle Lonicera japonica Thunberg. These two species often occur in dense monospecific stands which are separated by a strikingly narrow zone of mixing. This pattern of an introduced annual grass invading established stands of Japanese honeysuckle, a perennial woody vine which is known for its aggressiveness on fertile sites of the Carolinas (Bruner and Shearin, 1964), led me to ask the following questions: (1) Is M. vimineum invading undisturbed stands of L. japonica? (2) Does M. vimineum create a seedbank? (3) Can M. vimineum invade undisturbed stands of L. japonica if its seeds are sown in the stands in large numbers? (4) What environmental variables are related to success of M. vimineum sown in stands of L. japonica? (5) Does M. vimineum respond to an increase in essential soil elements with greater reproductive success? (6) Does M. vimineum alter the soil so as to inhibit growth of L. japonica? (7) What is the role of disturbance in the invasion of M. vimineum? Microstegium vimineum is in the tribe Andropogonae, subfamily Panicoideae. It was first collected in the United States in 1919 near Knoxville, Tennessee (Fairbrothers and Gray, 1972). By 1933 it had been collected in the mountains of western North Carolina (Blomquist, 1948; Fairbrothers and Gray, 1972) and by 1964 in 35 counties in North Carolina, primarily in the Piedmont (Radford et al., 1968). By 1972 the species had spread to at least 14 eastern states, from Florida to New Jersey and W to Ohio and Mississippi (Fairbrothers and Gray, 1972), and by 1978 to Arkansas (Smith, 1978). In Asia the species occurs in Japan, Korea, China, Ryukyus, Formosa, Malaysia, India and the Caucasus (Ohwi, 1984). Microstegium vimineum is unusual in that it is a C4 plant which, unlike typical C4 plants, is adapted to low light conditions (Brown, 1977, Winter et al., 1982). Under netting which transmitted 18% full sunlight, dry matter production was not significantly reduced from production in full sunlight; at 5% full sunlight, growth was 17% of that at full sunlight (Winter et al., 1982). Apparently, the C4 pathway itself has no inherent disadvantages, compared to the C3 pathway, under conditions of low light (Pearcy, 1983; Pearcy and Troughton, 1975).