Most KBS LTER research is carried out in a series of 11 types of plant communities, ranging from annual corn-soybean-wheat rotations to late-successional deciduous forest. All communities are replicated within the landscape. Our experimental design provides four annual cropping systems managed with a range of chemical-input intensities (from full to zero chemical inputs); two perennial cropping systems (one herbaceous [alfalfa] and the other woody [Populus sp.]); and two successional communities (one historically tilled and one never tilled). In 1993 we added three additional communities to the design, for a total of 5 unmanaged communities that now include three later successional oldfields abandoned from cropping 40-60 years ago, three planted conifer stands, and three older-growth hardwood stands. The design thus provides a wide range of replicated communities with the same pedogenic history that differ in key ecological characteristics (e.g. plant species diversity, productivity, litter quality, microclimate). This allows us to test specific hypotheses from which we can better infer the ecological mechanisms that confer productivity in row-crop ecosystems ? mechanisms that can then be tested with specific manipulative experiments. Baseline measurements are taken from all 11 community types, but not all communities are used to test every project hypothesis.