Since the ultimate success or a below-knee amputation depends upon preoperative, intraoperative and postoperative factors, it is not surprising that success or failure is difficult to predict on the basis of assessment of a single preoperative variable. In this study, discriminant function analysis was carried out on data obtained prospectively from 33 patients who had below-knee amputation. It showed that, by using the level of ankle systolic pressure and the severity of diabetes and of infection as discriminators, patients could be divided quite accurately into two groups - those whose amputation would be successful and those whose amputation would not. The study also showed that by using multivariate analysis to construct a single index of prediction, success can be forecast with a reasonably high degree of accuracy. Multivariate analysis is more accurate than the single-variate method.