Ambulatory monitoring and maximal treadmill exercise were compared in 40 normal subjects and 31 patients with mitral prolapse. A variable arrhythmia spectrum was observed in prolapse during monitoring: premature ventricular contractions in 18 (58%), supraventricular arrhythmias in 11 (35%), and bradyarrhythmias in 9 (29%). Significantly less arrhythmias occurred in normal subjects during monitoring: 10 (25%, P greater than 0.001), 3 (8%, P less than 0.001), 4 (10%, P less than 0.05), and 2 (5%, P less than 0.02), respectively. In patients with prolapse, arrhythmias occurred on resting electrocardiogram (ECG), 35% premature ventricular contractons, 6% supraventricular arrhythmias, and 10% bradyarrhythmias, and on treadmill exercise, 45%, 10%, and 3%; therefore, ambulatory monitoring was the most sensitive method of arrhythmia detection. No correlation existed between clinical features of prolapse and arrhythmias. Thus, arrhythmias occur in most patients with mitral prolapse, are not predictable by clinical characteristics, comprise a spectrum of ventricular and supraventricular tachyarrhythmias and bradyarrhythmias, and are best detected by ambulatory ECG monitoring.