Background: Spatial dispersion of atrial activation is a cause of postoperative atrial fibrillation (PoAF) after cardiac surgery. This study evaluated whether temporal dispersion of atrial activation causes PoAF after surgery in a clinical setting. Methods: Nineteen patients were enrolled. Postoperative atrial activation was evaluated by 24-hour Holter electrocardiography, with atrial pacing wires on the right atrium, for 5 days after cardiac surgery. No patient received antiarrhythmic drugs, including beta-blockers. The cycle length of 15 continuous atrial beats was measured at 4 time points: (i) earlier than 12 hours before PoAF, as a control, (ii) just before PoAF onset, (iii) during PoAF, and (iv) just before cessation of PoAF. Inhomogeneity of atrial activation was quantified by using the variation coefficient for a cycle length of 15 atrial beats during each phase. Results: The median inhomogeneity index of atrial activation (interquartile range) was 0.102 (0.046-0.136) in controls, 0.943 (0.582-1.610) just before PoAF onset (vs. control; p=0.009), 0.966 (0.631-1.117) during PoAF, and 0.471 (0.138-0.645) just before cessation of PoAF. Conclusions: Dispersion of atrial activation significantly increased just before PoAF onset. Temporal dispersion of atrial activation is a precursory variation of PoAF.