Homogeneity of recovery time protects against arrhythmias whereas dispersion of recovery time is arrhythmogenic. A single surface electrocardiographic QT interval gives no information on recovery time dispersion but the difference between the maximum and minimum body surface QT interval may be relevant. This hypothesis was tested by measuring the dispersion of the corrected QT interval (QTc) in 10 patients with an arrhythmogenic long QT interval (Romano Ward and Jervell and Lange-Nielsen syndromes or drug arrhythmogenicity) and in 14 patients without arrhythmias in whom the QT interval was prolonged by sotalol. QTc dispersion was significantly greater in the arrhythmogenic QT group than in the sotalol QT group. In patients with prolonged QT intervals, QT dispersion distinguished between those with ventricular arrhythmias and those without. This supports the hypothesis that QT dispersion reflects spatial differences in myocardial recovery time. QT dispersion may be useful in the assessment of both arrhythmia risk and the efficacy of antiarrhythmic drugs.