To determine whether detection of residual blockade is improved by using the accelerograph. A secondary objective was to compare acceleromyographic measurements with those obtained by electromyography.In a prospective, randomized, double-blind investigation, 22 anaesthetized children were studied during recovery from neuromuscular blockade following 0.1 mg.kg-1 vecuronium i.v.. Assessments of depth of block began 10 min after injection and were repeated at one minute intervals using electromyography (Datex, Relaxograph) in one hand, and acceleromyography (Biometer, Tofguard) in the other, to measure response of the adductor pollicis to train-of-four (TOF) stimulation of the ulnar nerve. Monitoring was stopped when no fade was visible and TOF ratio > or = 0.7. The electromyographic (EMG) and acceleromyographic (AMG) data were compared with corresponding observations of the number of twitches and TOF fade in the visible responses of the thumb, made by the attending anaesthetist. The method of Bland and Altman was used to compare differences between AMG and EMG data.During recovery from neuromuscular blockade, fade was no longer visible clinically 38.6 +/- 10.4 min (mean +/- SD) after the administration of vecuronium. This corresponded to TOF ratios of 0.04 +/- 0.23 by AMG and 0.34 +/- 0.21 by EMG. Usually, two twitches were visible before AMG detected the first twitch. The time to TOF ratio > or = 0.7 by AMG and EMG was similar at 49.1 +/- 10.5 and 50.9 +/- 9.0 min, respectively. The bias between AMG and EMG was one minute, with limits of agreement from -10 to nine min.AMG is superior to visual assessment in detecting residual neuromuscular block and provides similar estimates of recovering block as the more cumbersome EMG.