The effects of intraperitoneal administration of two different batches of human albumin (batch A and batch B) on peritoneal solute transport and dialysate white cell count were studied in 16 CAPD patients. The studies were done on two separate days during a 4-h dwell, one day without and one day with the intraperitoneal administration of 10 g/l human albumin. Marked differences were found between the two batches. The transport of all measured solutes increased during administration of batch A compared to the control experiments: urea 78% +/- 62%, lactate 51% +/- 38%, creatinine 96% +/- 54%, glucose 67% +/- 55%, inulin 27% +/- 33% IgG 126% +/- 80%, mean +/- SD; P less than 0.02). In the experiments with batch A the white cell count of the test bag was greater than that of the effluent ('night bag') before the test (13 +/- 5 vs 194 +/- 61 mm3/l; P less than 0.02). These effects were also observed when the dialysate was buffered to pH = 7.4 before inflow. Albumin batch B showed no effect on solute transport and white cell count. The effects of solute transport and white cell count are probably caused by the greater concentration of prekallikrein activator in batch A when compared to batch B (30.1 vs 0 U/l). Caution is warranted when human albumin is used for simultaneous measurement of peritoneal fluid and solute kinetics.