High-risk mammographic patterns may be used as a surrogate end point for breast cancer in etiologic research as well as in prevention studies. Physical activity may be one of the few modifiable risk factors for breast cancer. We examined the relationship between physical activity and mammographic patterns among 2720 Norwegian women, ages 40-56 years, who participated in both the Second and Third Tromso studies. Epidemiologic data were obtained through questionnaires. Two questions from the Second Tromso study and five questions from the Third elicited information on physical activity. The mammograms were categorized into five groups based on anatomical-mammographic correlations. For analysis, patterns I through III were combined into a low-risk group and patterns IV and V into a high-risk group. Odds ratios that were adjusted for age, education, menopausal status, body mass index, parity, age at menarche, oral contraceptive use, and alcohol intake, with 95% confidence intervals, were estimated using logistic regression. Women who reported moderate physical activity, i.e., more than 2 h/week, were 20% less likely (odds ratio, 0.8; 95% confidence interval, 0.6- 1.1) to have high-risk mammographic patterns compared with those who reported being inactive. This relationship remains consistent when stratified by menopausal status, parity, and tertiles of body mass index. However, all of the associations between various measures of physical activity and high-risk patterns found in this study are weak with confidence intervals that include 1.0. Thus, chance is a reasonable explanation for the weak associations found. The relationship between physical activity and high-risk patterns should be examined further as a means to explore the biologic mechanisms relating physical activity to breast cancer risk.