Question: Can patterns of species frequency in an old-field be explained within the context of a metapopulation model? Are the patterns observed related to time, spatial scale, disturbance, and nutrient availability? Location: Upland and lowland old-fields in Illinois, USA. Method: Species richness was recorded annually for seven years following plowing of an upland and lowland old-field subject to crossed fertilizer and disturbance treatments (mowing and rototilling). Species occupancy distributions were assessed with respect to the numbers of core and satellite species. Results: In both fields, species richness became higher in disturbed plots than in undisturbed plots over time, and decreased in fertilized plots irrespective of time. A bimodal pattern of species richness consistent with the Core-satellite species (CSS) hypothesis occurred in the initial seed bank and through the course of early succession. The identity of native and exotic core species (those present in > 90% of blocks) changed with time. Some core species from the seed bank became core species in the vegetation, albeit after several years. At the scale of individual plots, a bimodal fit consistent with the CSS hypothesis applied only in year 1 and rarely thereafter. Conclusions: The CSS hypothesis provides a metapopulation perspective for understanding patterns of species richness but requires the assessment of spatial and temporal scaling effects. Regional processes (e.g. propagule availability) at the largest scale have the greatest impact influencing community structure during early secondary succession. Local processes (e.g., disturbance and soil nutrients) are more important at smaller scales and place constraints on species establishment and community structure of both native and exotic species. Under the highest intensity of disturbance, exotic species may be able to use resources unavailable to, or unused by, native species. ?? IAVS; Opulus Press.