Two hypotonic but non-haemolysing irrigating solutions, sorbitol-mannitol (2% + 1%) and glycine (1.5%), were compared in 40 TURP cases using a continuous resection technique. Ethanol (1%) was added to the irrigating fluid as a marker to make possible early detection of fluid absorption by breath analysis. Mannitol and sorbitol were determined in plasma and urine; glycine and ethanol were determined in plasma. Apparent absorbed fluid volumes were calculated from the immediate postoperative plasma concentrations of ethanol, mannitol, sorbitol and glycine and from the elimination of mannitol in urine during 24 hours following the operation. The use of a continuous operating technique with a suprapubic trocar resulted in very small absorptions (less than 1 l) in this series. The concentrations of the two solutes in the sorbitol-mannitol irrigating fluid were balanced so that the plasma concentrations immediately postoperatively were of the same order when absorption occurred. The sorbitol concentration declined more rapidly than the mannitol concentration in conformity with previous findings. In most cases the peak plasma level was observed immediately postoperatively but in some cases at a later time (during the interval 0-2 hours), indicating absorption from a depot of fluid accumulated extravesically in addition to direct intravenous absorption. The best estimate of fluid absorption seems to be obtained from the urinary elimination of mannitol, followed by estimates based on the plasma mannitol concentration immediately postoperatively. The plasma ethanol level determined at the same time gave an estimate of the same order, whereas plasma sorbitol and glycine levels gave lower estimates (owing to rapid redistribution and metabolism).(ABSTRACT TRUNCATED AT 250 WORDS)