We have investigated the effects of 9.5% and 14.1% MAC concentrations of isoflurane on some psychometric measurements. Both concentrations depressed peak saccadic velocity (P < 0.01), choice reaction time (P < 0.05) and visual analogue scores for sedation (P < 0.05), but not the critical flicker fusion threshold. The incidence of errors in saccade tasks increased in a dose-related fashion, which made analysis of peak saccadic velocity less accurate at more than 10% MAC. The percentage error itself was an indicator of the depth of sedation. All the objective measures correlated highly with the estimated brain tension of isoflurane (r2 = 0.86-0.96), but not the visual analogue score for sedation (r2 = 0.51). This suggests that a combination of peak saccadic velocity, percentage error and choice reaction time is a potentially useful batch of tests to measure recovery from anaesthesia.