Eighty patients who presented for surgery of the forearm or hand were allocated randomly to one of two groups. In Group A, surgery was performed under supraclavicular brachial plexus block only; a mixture of equal parts of prilocaine 1% and bupivacaine 0.5% without adrenaline was used. In Group B, supraclavicular brachial plexus block was performed using prilocaine 1% alone, but in addition discrete nerve blocks were performed at elbow level using 0.5% bupivacaine without adrenaline. Patients in Group B had a significantly shorter duration of unwanted postoperative motor blockade and a significantly longer duration of postoperative analgesia (p less than 0.005).