Tourniquet-induced pain is probably mediated by C-fibres. The ability of morphine to interrupt this nociceptive conduction was studied in a double-blind fashion by administering either morphine 0.3 mg or saline intrathecally along with hyperbaric bupivacaine 15 mg for spinal anaesthesia in 40 patients undergoing orthopaedic surgery on the lower extremity. The block characteristics were similar in both groups. During surgery, no patient in the morphine group complained of pain, whereas in the saline group one patient complained of pain caused by the tourniquet and two other patients experienced surgical pain. A 60-min experimental thigh tourniquet pressure provocation (53 kPa for 20 min; 0 kPa for 20 min; 53 kPa for 20 min) on the contralateral unoperated extremity was then performed following surgery, when the spread cephalad of the sensory block had decreased below the T10 dermatomal level. Eight patients in the morphine group experienced no pain in this test, compared with two patients in the saline group (P less than 0.05). The remaining 12 patients in the morphine group had pain responses similar to those patients not given morphine. Intrathecal morphine provides a level of prophylaxis against tourniquet pain. However, the dosage employed here was associated with urinary and emetic side-effects.