A simple method was developed for recording epicardial monophasic action potentials intraoperatively in man. Potentials were recorded rapidly and reliably from multiple epicardial sites using a hand-held probe. The feasibility of repolarization mapping was assessed in 30 patients. It was possible to record technically adequate signals from most left ventricular and posterior right ventricular sites (success rates 72% and 61%, respectively), but it was more difficult to achieve satisfactory recordings over the free wall of the right ventricle (success rate 36%). The onset of cardiopulmonary bypass resulted in a transient rise in action potential duration of approximately 30-40 ms in the first minutes of bypass. An approximate steady state was achieved thereafter. Reproducibility of the method was assessed by mapping a small number of sites repeatedly. The 95% confidence limits for an estimate of repolarization time, based on a single potential, were +/- 21 ms. With 4 probe applications at each site, the confidence limits of the estimated site mean were reduced to +/- 8 ms. Based on this approach, statistically significant differences between sites were demonstrated in 9 of 10 patients, confirming that the method is sensitive enough to provide information on regional differences in repolarization. Intraoperative epicardial repolarization mapping has been shown to be a practical possibility. The technique holds promise as a means of assessing the role of dispersion of repolarization in arrhythmogenesis in man.