A follow-up study of 1,764 female patients, discharged alive from two tuberculosis sanatoria in Massachusetts between 1930 and 1954, was conducted. In the course of air collapse therapy of the lung (pneumothorax and pneumoperitoneum), 1,047 women were fluoroscopically examined an average of 102 times over a period of several years. A comparison group of 717 women with tuberculosis received other treatments that did not require fluoroscopic monitoring. A 1975 mailing address or a death certificate was obtained for 93.6% of the study subjects, and 78% of 1,146 living patients responded to a mailed questionnaire. An excess of breast cancer (41 observed and 23.3 expected) was seen among the fluoroscopically examined women, whereas no excess (15 observed and 14.1 expected) was apparent among the comparison women. When age was considered, the greatest absolute excess breast cancer risk occurred among exposed women who were first treated between the ages of 15 and 19 years. Among those women 30 years of age and older at the time of first exposure, no elevated breast cancer risk was detected; the failure to observe an excess, however, may be a matter of sampling variation inasmuch as a 50% increased risk could not be excluded. The concentration of excess breast cancer cases among women initially exposed before the age of 30 years suggests that young women may be particularly sensitive to radiation injury. Breast cancer excess risk was also seen to increase with increasing radiation dose, and a linear dose-response relationship is consistent with the data. The excess breast cancer risk did not appear until 15 years after initial exposure and was present at the end of 40 years of observation. The radiation risk estimate for the women living 10 years or more after first exposure is 6.2 excess breast cancer cases per million woman-year-rad.