In a prospective, randomized, double-blind study in 55 women undergoing elective caesarean section under spinal anaesthesia, we compared epidural diamorphine 3 mg (2 distinct boluses, group ED) with single-dose intrathecal morphine 0.2 mg (group SM), in terms of analgesic efficacy, patient satisfaction and side-effects at 2, 3, 4, 8, 12, 16, 24 and 28 h postoperatively. There were no significant differences between groups in pain (assessed by 100 mm visual analogue scale), incidence of pruritus, sedation or respiratory depression measured by continuous pulse oximetry. However, time to first request for supplementary oral analgesia was longer in SM than in ED (mean +/- SD: 22.3+/-12.0 h vs. 13.8+/-6.5 h, P = 0.04). The incidence of nausea or vomiting was significantly higher in SM than ED (73% vs. 41%, P = 0.01). In ED, the mean +/- SD time to requirement of the second bolus was 6.7+/-3.2 h. There was a high level of satisfaction in both groups. We conclude that two boluses of epidural diamorphine 3 mg and single-dose intrathecal morphine 0.2 mg provide satisfactory analgesia after caesarean section, but spinal morphine was associated with both delayed requirement for supplementary analgesia and a higher incidence of nausea and vomiting.