Although exposure to ionizing radiation is a recognized risk factor for breast cancer, the potential hazard from low-dose, fractionated exposures during early breast development has not been thoroughly evaluated. Women with scoliosis represent a valuable population for studying this issue because they are exposed to multiple diagnostic x rays during childhood and adolescence, times when the breast may be highly sensitive to the carcinogenic effects of radiation. A study was conducted of 1,030 women with scoliosis who were seen at four Minneapolis area medical facilities between 1935 and 1965. The average age at diagnosis was 12.3 years; 60% of the women had idiopathic scoliosis. Individual x-ray films were counted and the number per patient ranged from 0 to 618 films (mean, 41.5). On average, the x-ray exposures were given over an 8.7-year period. Ninety percent of the women were located, of whom over 92% responded to a mail questionnaire or telephone interview. The average period of observation was 26 years. Overall, 11 cases of breast cancer were reported, compared with six expected (standardized incidence ratio = 1.82, 90% confidence interval = 1.0-3.0). Excess risk increased with time since exposure and was highest among those followed for more than 30 years (standardized incidence ratio = 2.4). Risk also increased with the number of x rays and with the estimated radiation dose to the breast (mean, 13 rad). These data suggest that frequent exposure to low-level diagnostic radiation during childhood or adolescence may increase the risk of breast cancer.