Following mastectomy, patients aged 35 to 76 years with operable breast cancer underwent postoperative irradiation of the chest wall and regional lymph nodes. They were then assigned at random to receive no further therapy (NT), ovarian irradiation (R) 2000 rads in 5 days, or ovarian irradiation in the same dosage plus prednisone (R + P) 7.5 mg daily for up to five years. A total of 703 eligible patients received the randomly assigned treatment. The median follow up was 21 years with a range of 14 to 25 years. Overall, there was a delay in recurrence (p = 0.03) and survival was prolonged (p = 0.19) for patients who received R, but in neither case was the difference significant after adjusting for the multiplicity in our data. Overall, patients who received R + P experienced a significant delay in recurrence (p = 0.0003) and a significantly prolonged survival (p = 0.005), even after adjusting for multiple comparisons. In premenopausal patients who received R, the recurrence of breast cancer was delayed and survival prolonged, but not significantly. In premenopausal women aged 45 years or more, R + P therapy significantly prolonged survival (p = 0.0004), while the delay in recurrence although significant (p = 0.02) was only marginally so after allowance for multiple comparisons. No value was demonstrated for ovarian irradiation with or without prednisone therapy in postmenopausal patients. A new finding in this review was that contralateral breast cancer as the first failure was reduced by R + P compared to NT in the overall group.