Concern over undesirable environmental impacts and low profitability of conventional potato (Solanum tuberosum L.) production systems has led to efforts to reduce agrichemical use while improving yield and net returns. Weed management based on a better understanding of weed responses to tillage and other cropping practices is a key to success in these efforts. We conducted a 4-yr field experiment on a Caribou gravelly loam (coarse-loamy, mixed, frigid Typic Haplorthods) in Presque Isle, ME, in which potato was grown in alternating years with either oat (Avena sativa L.) or berseem clover (Trifolium alexandrinum L.), in plots that were either moldboard plowed (MP) or chisel plowed (CP). To determine yield losses due to weed interference, potato was grown either in weed-free conditions or with weeds that survived cultivation and billing. Seed density, plant density, and biomass of dominant weed species (Chenopodium album L. and Brassica and Raphanus spp.) were lower following MP than CP. Average loss of U.S. No. 1 tubers due to weed interference was 12% with MP, compared with 43% with CP. Weed and potato responses to rotation crops differed between years. Weed biomass production was lower in oat than in berseem clover during the 1990 rotation crop phase. In the subsequent 1991 potato crop, yield loss due to weed infestation was 13% following oat, compared with 38% following berseem clover No difference in weed growth was evident between oat and clover rotation crops in 1992, and there were no rotation crop effects on weed-related yield loss in the subsequent 1993 potato crop. me present a conceptual model that incorporates these data, links need life stages and crop performance, and illustrates the effects of interference between weed species. The model suggests that, if need growth in rotation crops can be strongly suppressed, chisel plowing should result in low densities of annual weed species and minimal yield loss in subsequent potato crops.