Thirty patients undergoing upper laparotomy were entered into a randomized trial, comparing the effect of midthoracic (T) and lumbar (L) epidural morphine on postoperative pain and pulmonary function. Five mg morphine was injected through the catheter at the end of the operation, and subsequently three times a day. Six, 30 and 54 h postoperatively, the following tests were performed: linear analogue pain score, arterial gas tensions (PaO2, PaCO2 and pH), forced ventilatory capacity (FVC), forced expiratory volume in 1s (FEV1) and peak expiratory flow rate (PEF). The changes in pain score (increase of the median): T: 21, 6, 5, and L: 24, 15, 8 per cent of full scale), PaO2 (decrease of the tension: T: 1.7, 2.1, 2.4, and L: 2.0, 2.8, 2.0 kPa), PaCO2, pH, FVC (decrease of the volume: T: 1.3, 1.1, 0.9, and L: 1.3, 1.3, 1.21), FEV1 and PEF from the preoperative tests were not significantly different. It is concluded that the clinical effect of epidural morphine for postoperative pain treatment is the same or little different whether the administration takes place at the thoracic or lumbar level.