A retrospective analysis of 73 patients treated for primary vaginal carcinoma with radiation therapy was performed to evaluate the effect of radiotherapeutic technique on local control. Local control was achieved in five of 22 patients (23%) treated with pelvic external beam therapy alone, three of four patients (75%) treated with intracavitary cylinder or Bloedorn applicator alone, and 30 of 47 patients (64%) treated with combination of external beam and brachytherapy. Radiation therapy complications requiring hospitalization occurred in six patients (8%). A statistically significant difference in local control was achieved only when patients receiving external beam and brachytherapy were compared with patients receiving external beam therapy alone (P < 0.005). Total mid-tumor dose was defined as the sum of midplane tumor dose from external beam therapy, mid-tumor dose from interstitial radium needles, and the vaginal surface dose from intracavitary radium systems. Total mid-tumor doses ranged from 16 to 121.7 Gy. Only two of 16 patients receiving less than 55 Gy total mid-tumor dose achieved local control. As a result, dividing doses of 45, 55, 65 and 75 Gy produced a statistically significant superior local control rate in the patients receiving the higher dose (P < 0.01). None of the 16 patients receiving less than 55 Gy total mid-tumor dose had received brachytherapy. We conclude that the combination of external beam therapy and brachytherapy is essential to achieve optimal control of primary vaginal carcinoma.