During recent years, the exploration of different aspects of atrial fibrillation (AF) has become increasingly interesting. Thus, knowledge about basic underlying mechanisms, consequences and different modes of treatment has rapidly expanded. At a meeting in Lund, Sweden, in 1993, scientists within different fields of AF research gathered for the exchange of information. This paper is a short summary of some topics discussed at the Lund meeting and some suggestions as to how further research in this field may help to improve our understanding of this arrhythmia and the treatment of patients suffering from it. Underlying pathoelectrophysiological mechanisms in AF have been explored in experimental models in animals and by direct recordings of different atrial myocardial electrophysiological variables both in the catheter laboratory and during open heart surgery in man. Some findings illustrate possible generalized atrial myocardial mechanisms, whilst other findings clearly indicate the possibility of localized pathoelectrophysiological mechanisms. The generally accepted hypothesis that AF is perpetuated by multiple re-entry mechanisms is, thus, both verified and modified by recent studies. In addition to subjective symptoms and well identified thromboembolic consequences, accumulating evidence tells us that AF may precipitate a myocardial dysfunction which may be misinterpreted as an underlying factor initiating the arrhythmia. Today's treatment of AF includes several newer antiarrhythmic drugs, different ablation techniques, the application of different electrical devices as well as different surgical methods. New, improved and simplified methods are expected. Atrial fibrillation is the single most important supraventricular arrhythmia needing substantial further exploration of mechanisms, consequences and treatment. The Lund symposium contributed to this process by defining the state of knowledge in 1993 and outlining the need for the years to come.