To investigate the controversy surrounding the life-expectancy of patients on continuous ambulatory peritoneal dialysis (CAPD) compared with that of patients on haemodialysis or transplantation mortality data from 389 patients accepted for renal replacement therapy in Leicester between July, 1974, and July, 1985, were retrospectively analysed with respect to a wide range of pre-treatment variables (6 scales and 115 binary variables), by a method (Cox's) that adjusts for the distorting influence of selection bias. 9 independent variables were identified as having a significant influence on survival. Adverse factors were age, amyloidosis, ischaemic heart disease, convulsions, and acute presentation. Beneficial variables were male sex, parenthood, pyelonephritis, and residence in Leicestershire. By correcting for the influence of these variables and using time-dependent treatment co-variates, the bias adjusted estimates of the relative risk of death were 1 for patients on CAPD, 1.30 for those on haemodialysis, and 1.09 for patients who received transplants. These risks do not differ significantly from one another and suggest that CAPD is at least as effective as haemodialysis or transplantation at preserving life.