In a screening program of self-referred women, different mammographic parenchymal patterns were related to significantly different rates for developing breast cancer. The risk of cancer detection subsequent to a negative mammographic examination was 7.6 times greater for women in the highest parenchymal risk class compared with the lowest, an increase in risk comparable to that of a personal history of breast cancer and greater than that reported for any other combination of historical risk factors. These differences are qualitatively similar to, but of a lesser magnitude than, those in previous reports which were based on symptomatic women with previous negative mammograms. Data suggest this difference in risk is inherent between parenchymal patterns, rather than indicating difficulty in identifying small cancers in dense breasts. Findings of differential parenchymal risk, coupled with other risk factors, may lead to concentrating mammographic screening on a smaller segment of the population, thus improving the benefit-to-cost ratio.