Studies of 2216 24-hour Holter electrocardiograms of 1549 patients revealed bradyarrhythmias in 33 cases of advanced atrioventricular block, 169 cases of sick sinus symdrome and 32 cases of atrial fibrillation with slow ventricular response. The underlying diseases were mostly idiopathic, but included ischemic heart diseases and cardiomyopathy as well. The findings of Holter electrocardiography showed a significant difference between the bradyarrhythmia group and control group as to the minimum heart rate, maximum heart rate, the longest ventricular pauses and total number of QRS's during a 24-hour period. There were 55 cases of bradyarrhythmias with longest ventricular pauses of 3.0 seconds or longer. The incidence of atrial fibrillation with slow ventricular response was significantly different from other bradyarrhythmias. There were 64 cases of bradyarrhythmias with the total number of QRS's during a 24-hour period equaling 70,000 beats/day or less. The incidence of sick sinus syndrome group I was significantly different from other bradyarrhythmias. As regards severity of bradyarrhythmias, these two findings were inconsistent. Comparative studies of paced and unpaced groups showed that the patients were older and had more subjective symptoms in the paced group. In all Holter findings, the total number of QRS's during a 24-hour period and the longest ventricular pauses were significantly different from other findings. No correlation was observed between the longest ventricular pauses and the maximum sinus node recovery time (or maximum automaticity recovery time). In consideration of subjective symptoms and underlying diseases, severity and pacemaker indication of bradyarrhythmias should be determined by longest ventricular pauses and total number of QRS's during a 24-hour period shown on 24-hour Holter recordings.