Weed suppression by cover crops grown during the winter fallow period in continuous corn may lead to a reduction in herbicide use. Rye, crimson clover, and subterranean clover cover crops were compared with corn stubble under a conventional management system (CS) that included plowing and use of preemergence residual herbicides and a low-input management system (LIS) that included no-tillage and use of a presowing nonresidual herbicide for three consecutive years (1994–1996). Cover crop and above-ground weed biomass prior to desiccation were not influenced by management system. Cover crop biomass ranged from 1,420 to 5,657 kg ha−1 for rye, from 563 to 4,217 kg ha−1 for crimson clover, and from 563 to 4,248 kg ha−1 for subterranean clover. At crop planting, rye reduced weed biomass from 54 to 99%, crimson clover from 22 to 46% (with a negative value in 1995), and subterranean clover from 21 to 67%. Weed growth suppression was usually higher in years when cover crop biomass was higher. There were no differences in weed suppression by cover crops later in the season (corn in the fourth leaf stage), while total weed density was higher in LIS than CS in 2 of 3 yr. Total weed cover at corn's ‘full dent’ stage ranged from 1 to 7% in CS and from 24 to 47% in LIS. Cover crops influenced weed composition only in years when cover crop growth was high; otherwise their effect was masked by that of the management system. Weed communities showed higher diversity under LIS than under CS. Consistency of associations between weed species and treatments over sampling dates and years was found especially for some of the species associated with LIS. After 3 yr, redroot pigweed, common lambsquarters, and black nightshade were regularly associated with rye-LIS at an early corn growth stage; this may indicate a species shift toward a more troublesome composition.