Between 1976 and 1982, 293 patients were treated for carcinoma of the uterine cervix at Washington University by definitive radiotherapy consisting of external beam therapy and two standard Fletcher-Suit applications (tandem plus vaginal colpostats). In ninety-nine patients (34%) mini-colpostats (MC) were used for one or both of their intracavitary insertions while 194 (66%) patients were treated twice with regular Fletcher-Suit colpostats (RC). The frequency of MC use was related to the age and parity of the patients. The distribution by stage of MC and RC groups was not significantly different. Pelvic failure in the MC group was similar to that of the RC group (21% vs 24%). Five-year disease-free survival was also similar between the two groups: 86% vs 80% Stage IB, 57% vs 61% Stage IIA, 47% vs 52% Stage IIB, and 27% vs 45% Stage III for MC and RC groups, respectively. The rate of major complications (grade 3) was 15% in the MC group and 8% in the RC group (p = 0.08). Careful phantom dosimetric studies in both types of colpostats and correlations of dose distributions at various points in the pelvis with frequency of rectal and bladder complications were carried out. The bladder and rectum received a 5-10% higher mean radiation dose (Gy) in the MC group than in the RC group despite lower overall exposure (milligram-hours). Thermoluminescent dosimetry in a polystyrene phantom demonstrates that approximately 10% higher doses are delivered to the bladder, rectum, and point A with an MC system as compared to an RC system, for constant exposure in mgh. Phantom measurements of a newer MC with bladder and rectal shielding demonstrate no influence on the bladder and rectal point dose at a source separation of 3 cm; midline points of the bladder and rectum are not within the full shadow of the shields even if the colpostats are flush with the tandem. Implications for therapy are discussed.