In 1958, a co-operative clinical trial, involving 23 institutions, was begun to evaluate the efficacy of a short course of triethylenethiophosphoramide as an adjuvant to radical mastectomy in the treatment of patients with primary carcinoma of the breast. Of 826 women prospectively randomized so that they received either a placebo or triethylenethiophosphoramide, 99.3 per cent were available for five years of follow-up study, and 95.3 per cent have been observed for ten or more years. Results have indicated the inadequacy of standard operative therapy in effecting a permanent tumor-free state in a majority of patients. Particularly distressing was the finding that 76 per cent of all patients with positive axillary nodes--65 per cent with one to three and 86 per cent of those with more than or equal to four positive--demonstrated a treatment failure by ten years and that only 24.9 per cent survived. The survival rate of those with one to three positive nodes was 37.5 per cent and only 13.4 per cent if more than or equal to four nodes contained tumor. Also disturbing was the observation that one of four patients with negative axillary nodes displayed a treatment failure by ten years. Such findings mandate the urgency for immediate evaluation of available systemic therapeutic agents postoperatively. That systemic chemotherapy can enhance the disease-free state as well as survival rate of at least some patients with carcinoma of the breast is supported by the finding that the original observation at 18 months indicating an advantage for premenopausal women with more than or equal to four positive nodes who received triethylenethiophosphoramide persisted for as long as ten years. At that time, 21 per cent fewer patients in the triethylenethiophosphoramide group had treatment failures and 21 per cent more survived, an indication that the initial delay in treatment failure was reflected in patient survival rate. Data suggesting that the limited chemotherapy used was more effective in patients having smaller tumors are in keeping with the concept that adjuvant therapy is likely to be most effective in a host with minimal residual tumor burden. Of importance was the observation that 80 per cent of treatment failures occurring in ten years were apparent by five years of observation. Strikingly, 86 per cent of ten year treatment failures in patients with positive nodes occurred by five years. In patients with more than or equal to four positive nodes, this was so in 92 per cent of the instances. Such findings tend to discredit the claim that the worth of an alternative treatment for carcinoma of the breast can be ascertained only by a period of observation much longer than five years.