In nine patients with the tachycardia-bradycardia syndrome, a dysfunction of sinusal automatism was observed, shown by a prolonged recovery time of sinus node for six of the patients, and by an insufficient response to atropine in all nine. In two of the five patients, where a study of seno-atrial conduction was possible a conduction defect at this level was noticed. In three patients (33%), asystole was observed: in one patient in the course of a tachycardial crisis; during the passage of an atrial fibrillation in sinusal rhythm in another, and following premature isolated atrial pulsations in the third. The observations of asystole pause in the third patient following premature isolated beats which were longer than those induced electrically, led to the conclusion that, more than a sinusal automatism or sino-atrial conduction defect, asystole can be due to repetitive re-entries that are not reproduced at the level of the sino-atrial junction. The atrial response obtained in another patient with atrial stimulation slightly superior to the threshold, seems to exclude an atrial inexcitability hypothesis. The paroxysmal tachycardia and the atrial fibrillation were the most frequently noted arrhythmias; atrial flutter was only rarely observed. A rotation of rhythm disturbances was recorded in four of the nine patients, both on diverse occasions and during one crisis. The atrial electrostimulation performed on a patient during asystole pauses occurred spontaneously during the course of a tachy cardial crisis, allowed conduction of the atria only for 1-2 stimuli, and did not impede the recovery of the tachycardial paroxysm. It is probable that the conduction disturbances and/or the atrial excitability, which have the same determining cause of sinusal dysfunction, can be responsible for atrial arrhythmias, with the characteristic symptom of the tachycardia-bradycardia syndrome.