Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia and the major cardiac cause of stroke. Recent studies in patients with paroxysmal AF have shown that the arrhythmia is triggered by focal sources localized usually in one of the cardiac veins. However, in chronic AF, the prevailing theory is that multiple random wavelets of activation coexist to create an unorganized atrial rhythm. Experiments in isolated hearts have demonstrated that stable, self-sustained rotors can exist in the atria and that high frequency activation by such rotors results in the complex patterns of activation that characterize AF. Studies in animals and patients support the view that at least some cases of paroxysmal and chronic AF are the result of the uninterrupted periodic activity of discrete reentrant sites. In this brief review article, we examine historical data and more recent experimental evidence behind the hypothesis that AF may be organized by one, or a small number of high-frequency reentrant sources localized in the left atrium. We then discuss the potential implications and evidence supporting such a hypothesis for human AF. Finally, we suggest future studies designed to unravel the detailed molecular, cellular and pathophysiological mechanisms responsible for AF initiation and maintenance. The work discussed may open potentially exciting new diagnostic and therapeutic possibilities.