Epidemiological evidence has generally supported a protective association of physical activity with large-bowel adenomas, but whether the protective effects are restricted to recent or past activity is uncertain. We determined whether recent and past recreational or total daily activity was associated with prevalence of colorectal adenomas among male and female members of a prepaid health plan in Los Angeles who underwent sigmoidoscopy (n = 488 matched pairs). Participants, aged 50-74 years, completed a 126-item semiquantitative food frequency questionnaire and were also interviewed regarding non-dietary risk factors in 1991-93. In the univariate analysis, all measures of recent recreational physical activity were associated with reduced prevalence of polyps. After adjustment for body mass index, smoking status, daily servings of fruit and vegetables, use of non-steroidal anti-inflammatory agents and intakes of calories, saturated fat and alcohol, the associations were weakened. For subjects engaging in high-intensity activities compared with subjects not engaging in vigorous activities, the multivariate odds ratio (OR) for recent recreational activity was 0.7 [95% confidence interval (CI) 0.4-1.1, trend P = 0.08]. Past recreational activity and past or recent total daily activity were not associated with prevalence of adenomas. These results support a modest association of recent recreational physical activity with prevalence of colorectal adenomas.