Dose-response relationship of mini-dose intrathecal morphine (0.025-0.125 mg) for analgesia after Cesarean section was studied. Sixty-three patients were randomly divided into six groups by the following intrathecal morphine injection: group 1 (0 mg), group 2 (0.025 mg), group 3 (0.05 mg), group 4 (0.075 mg), group 5 (0.1 mg), and group 6 (0.125 mg). The selected dose of morphine mixed with 2 ml 0.5% hyperbaric bupivacaine (10 mg) was administered intrathecally to induce spinal anesthesia. The mean analgesia duration in each group was 3.6 +/- 2.0, 10.6 +/- 7.1, 17.3 +/- 13.8, 25.6 +/- 7.5, 33.9 +/- 10.1, and 39.5 +/- 11.9 h respectively (mean +/- SD). In morphine groups, duration of analgesia was significantly longer (p less than 0.05) than control group (0 mg), and the first 24 h pain scores were also lower (p less than 0.01). Furthermore, a significant linear dose-response relationship between analgesic duration and the dose of intrathecal morphine was revealed (y = 3.28 + 295.5x, r2 = 0.64, p less than 0.05). Among morphine groups, analgesic quality was significantly better in patients in groups 4-6 than those in group 2 and 3 (p less than 0.05), so as in the proportion of effective analgesia in the first 24 h (p less than 0.01). Neonatal condition was not adversely affected by such mini-dose of intrathecal morphine. The most common maternal adverse effect observed was pruritus, and its incidence was significantly greater in groups 3-6 than in the control group (p less than 0.05). However, no significant difference was observed among all morphine groups.(ABSTRACT TRUNCATED AT 250 WORDS)