Although treatment results for Stage IB-IIA-B carcinoma of the cervix are generally stratified only by stage, several authors have found that other parameters such as tumor size, patient age, histology, and pretreatment hematocrit may influence pelvic control and/or survival. This is an analysis of 306 patients with carcinoma of the cervix treated with irradiation alone at the University of Florida between October 1964 and June 1984. Results were analyzed for the end points of pelvic control, distant metastasis, relapse-free survival, and overall survival. Stage for stage, there was a decrease of at least 20 percentage points in pelvic control rates of tumors greater than or equal to 6 cm versus those of tumors 0 to 3 cm in diameter, by univariate analysis; a similar pattern was also observed for the end point of relapse-free survival. When patients were stratified by Stages IB-IIA (greater than or equal to 6 cm and less than 6 cm) and IIB (greater than or equal to 6 and less than 6 cm), there was a clear pattern of improved pelvic control and relapse-free survival for patients over 50 years old and for patients with pretreatment hematocrits greater than 40%. The parameters of stage, tumor size, patient age, histology, and pretreatment hematocrit were evaluated in a multivariate analysis. For the end points of pelvic control, distant metastasis, and relapse-free survival, only tumor size was of independent prognostic significance for all three end points with p-values of 0.022, 0.003, and 0.0006, respectively. Stage did not show independent prognostic significance for any of these end points with p-values of 0.257, 0.878, and 0.284, respectively. The data suggest that tumor size is an important prognostic factor and should be incorporated into the reporting of cervical cancer treatment results.