BACKGROUND Intra-arterial regional anaesthesia (IARA) may be useful for ambulatory hand surgery in patients with poor veins. This randomized, double-blind study assessed which of the three doses of lignocaine gives the optimal analgesia with a minimum of adverse effects. METHODS A preservative-free, alkalinized 0.5% lignocaine 1, 2 or 2.89 mg/kg body weight was injected into the radial arteries of 60 adult patients, allocated to three equal groups, to produce anaesthesia for carpal tunnel releases, capsulotomies, tenosynovectomies, palmar fasciectomies, Z-plastics, arthroplastics, arthrodeses etc. RESULTS Surgical analgesia and motor block were best in group 3 (P < 0.01), whereas injection and tourniquet pain scores were similar in the three groups. Onset of analgesia was similar in all groups, and varied between 2 and 15 min. Cannulation time, surgery start time and tourniquet time were also similar in all groups, as were operating conditions and patient's acceptance of the method. No significant cardiovascular changes were observed after tourniquet release in any of the groups. Plasma lignocaine concentrations were lowest in group 1 (1 mg/kg) (P < 0.001). Five patients in group 1, seven in group 2 and seventeen in group 3 developed small bruises at the cannulation site (P < 0.001). Six patients (two in group 1, three in group 2 and one in group 3) had minor symptoms of lignocaine toxicity after tourniquet release (NS). No other complications were observed. CONCLUSIONS The highest dose of lignocaine produces best surgical analgesia, without increasing the risk of toxicity. However, many patients receiving this dose will develop bruises at the injection site, and an occasional patient may need supplemental analgesia.