We have used data from a large population-based case-control study in the United States to evaluate the effect of occupational physical activity on breast cancer risk. Women diagnosed with breast cancer identified from four state cancer registries, and controls randomly selected from lists of licensed drivers or Medicare beneficiaries, were interviewed by telephone for information on usual occupation and other factors. We classified usual occupation into one of four categories of physical activity. After excluding subjects for whom a strength rating could not be assigned, we had a final sample size of 4,863 cases and 6,783 controls. Using conditional logistic regression models, we calculated adjusted odds ratios (OR) and 95 percent confidence intervals (CI) for occupations having light, medium, and heavy activity compared with sedentary ones. Women with heavy-activity occupations had a lower risk of breast cancer than women with sedentary jobs (OR = 0.82, CI = 0.63-1.08), as did women with jobs with medium activity (OR = 0.86, CI = 0.77-0.97) or light activity (OR = 0.92, CI = 0.84-1.01). There was a significant decreasing trend in the ORs from sedentary to heavy work (P = 0.007). Although limited by exposure misclassification, these data are consistent with the hypothesis that physical activity reduces the risk of breast cancer.