This prospective, randomised, observer blinded study compared the onset time of brachial plexus block using 2% lidocaine 25-30 ml with adrenaline 5 microg.ml(-1) into the 'corner pocket' inferolateral/lateral to the subclavian artery (supraclavicular, n = 30) or to a triple point injection around the axillary artery (infraclavicular, n = 30). Mean (SD) onset time for complete pinprick sensory blockade assessed by a blinded observer in all four distal nerves was similar in both groups: supraclavicular = 22 (9.4) min, infraclavicular = 21 (7.1) min, p = 0.59. Complete sensory blockade in all four nerve territories at 30 min was achieved in 57% in group supraclavicular and 70% in group infraclavicular (p = 0.28). Painless surgery without the requirement for block supplementation was higher in group infraclavicular (28/30, 93%) compared with group supraclavicular (19/30, 67%; p = 0.01). Of the 11 failures in group supraclavicular, nine were due to incomplete ulnar nerve territory anaesthesia. These results do not support the concept of rapid onset successful supraclavicular block via a simple ultrasound-guided local anaesthetic injection inferolateral to the subclavian artery.