The nutritional component during childhood and adolescence that contributes to variation in adult height may be associated with the subsequent risk of breast cancer in women. Municipalities in three Norwegian counties were ranked and divided in three categories according to resident women's height, where variation in height is assumed to reflect variation in nutritional living conditions during childhood and adolescence. For each county observed number of incident cases (1975-1984) and deaths (1966-1987) of breast cancer in "short", "medium" and "tall" municipalities were compared with the expected number computed from the national rates. An analogous analysis was done, comparing incidence and mortality between 18 Norwegian counties, but using county-specific infant mortality for the period 1921-1935 as an alternative indicator of living conditions. There was a consistent positive association between height in municipalities and breast cancer incidence (overall P trend = 0.02) and mortality (overall P trend = 0.05). For infant mortality between counties, there was no association with breast cancer. In populations where variation in height reflects variation in nutritional living conditions during childhood and adolescence, there may be a positive association between height and risk and mortality of breast cancer.