The relation of Wolfe's parenchymal patterns and radiographically-assessed breast size with breast cancer risk was evaluated in a population-based nested case-control study in Uppsala, Sweden. All women who attended a mammographic screening programme in Uppsala county starting in 1988 have been followed for the occurrence of breast cancer through 1993. The analysis was based on 295 cases and 589 age-matched controls, whose mammograms were blindly evaluated for parenchymal pattern and breast size. Women with P2 or DY pattern had a significantly elevated risk of breast cancer compared with women with N1 or P1 (OR = 2.09; 95% CI = 1.52-2.86). There was an inverse association of breast size with breast cancer risk, which disappeared after adjusting for parenchymal pattern, because breasts of smaller size tended to have high-risk parenchymal patterns. It is concluded that in Swedish women, and perhaps in Caucasian women in general, small breast size is associated with increasing breast risk through its association with high-risk parenchymal pattern. This is in contrast to the fact that Asian women, who in general have breasts of smaller size, have low prevalence of high-risk parenchymal pattern as well as low rates of breast cancer.