This study was designed to test if changes in the degree of respiratory sinus arrhythmia (RSA) can be used as an index of light anaesthesia. An on-line, real-time data logging system was used to record simultaneously the EEG and ECG waveforms from 10 patients undergoing routine surgery using i.v. propofol (Diprivan) anaesthesia. The degree of RSA was determined in real-time. The median frequency of the EEG was derived off-line and correlated with the degree of RSA. Time series analysis was performed on the derived indices off-line. Significant changes in the degree of RSA occurred in response to changes in propofol infusion in all patients; these changes corresponded also to changes in the median frequency of the EEG. Heart rate and ventilatory frequency are measured routinely during surgery and it is suggested that on-line monitoring of RSA, derived from these standard signals, provides a more convenient and objective index of lightening anaesthesia than either EEG analysis or classical estimates of anaesthetic depth based on arterial pressure or heart rate.