Three patterns of care outcome surveys in prostate cancer totalling 1516 patients had been combined and analyzed for the effect of dose on infield recurrence. There are significant dose effects observed in the overall data (1516 patients, p = .003), Stage B cancers (725 patients, p = .004) and Stage C cancers (624 patients, p = .059). No dose effect was observed for Stage A cancers (168 patients, p = .217) within the dose range observed (5500 cGy to greater than 7000 cGy). For patients with Stage B cancer one may conclude that dose between 6000 cGy and 6999 cGy is appropriate. Patients treated to less than 6000 cGy show a highly significant increase in local failure. Patients treated to greater than 7000 cGy do not show a demonstrable improvement in local control, but do show an increase in complications. Patients with Stage C cancer appear to require dose that is equal or greater than 7000 cGy to obtain the best local control, and the potential increased morbidity of these high doses appears to be justified in this stage of the disease. Patients who have been given hormonal therapy more than 1 month prior to radiation therapy show an increase in local failure rate for all stages of cancer. This is presumed to be the selection of poor risk patients for adjuvant hormonal treatment or by referring non-responding hormone treated patients for radiation therapy. Histologic grade exerts a major influence on local failure for patients with Stage C disease (p = less than .001), identifying an important stratification point for prospective clinical trials and a sub-group for which it is important to develop strategies for improving local control. The policy of treating all stages of prostate cancer with the same dose is not supported by these data.