During the twenty two years, Aug. 1963 through Aug. 1985, we performed 83 cases of epiphysiodesis on patients with leg length discrepancy due to poliomyelitis. In an attempt to determine the accuracy of predicted correction, we analyzed predicted final discrepancy at the time of operation and actual final discrepancy at or near skeletal maturity. Of the 83 cases, 30 cases were available for final evaluation. Among four methods of prediction that have been employed, those of White and Stubbins, and Anderson, Green and Messner were statistically reliable. The method of Gill and Abbott was statistically not reliable in our series, presumably because it utilizes percentile distribution of stature based on Caucasian population in the United States. Mosley's prediction was also made on 8 of the cases. However, we were unable to reach statistically significant conclusion, obviously because of short period of preoperative assessment on limited number of cases at this time. Prediction of correction on 5 cases in which both lower extremit.ies were affected, was statistically not reliable in all methods employed.