The relation between plasma concentration and the effects of atracurium was studied in seven patients anaesthetised with thiopentone, fentanyl and nitrous oxide-oxygen. The response to train-of-four stimulation at ten-second intervals with tetanic stimuli applied every five minutes were recorded. The first sign of transmission returning after complete blockade was usually the post-tetanic facilitated twitch, which was noted when the mean atracurium concentration was 1.15 mg 1(-1) (SD 0.77). The most sensitive parameter was the train-of-four ratio, which recovered to 0.5 when the concentration was 0.217 mg 1(-1) (SD 0.56), compared with a concentration of 0.271 (SD 0.85) at 50% recovery for twitch height, 0.221 (SD 0.029) for tetanic peak and 0.231 (SD 0.079) for tetanic fade. The value for the train-of-four ratio differed significantly (P less than 0.05) from that for the twitch height, but other differences were not significant. Once recovery commenced, these four parameters recovered at similar rates, with recovery indices (25 to 75% responses) of 14.8 (SD 1.7), 14.0 (SD 1.0), 14.3 (SD 1.8) and 13.7 (SD 2.1) minutes respectively. Post-tetanic facilitation was most marked during severe but incomplete blockade and tetanic stimulation temporarily reversed the atracurium-induced decrease in train-of-four ratio. Clinically, the use of tetanic stimulation did not improve the sensitivity of neuromuscular monitoring, but post-tetanic count may be useful where monitoring of profound relaxation is required.