The management of patients with bacterial endocarditis complicated by atrioventricular block is based on uncontrolled data, mostly from retrospective surgical and autopsy series. It is difficult to advance broad recommendations on the basis of such a biased population. Nevertheless, it is the firm opinion of many experienced clinicians that heart block developing as the result of aortic endocarditis signals myocardial abscess formation, and thereby is an indication for early surgery. I present a patient with aortic and mitral endocarditis in whom first degree heart block developed and then disappeared over five days; she was successfully managed with medical therapy alone. This case illustrates that some patients with endocarditis and heart block will not require surgery. In this setting, I propose the following guidelines in selecting patients for operation: 1) the observed appearance or progression of heart block; 2) the presence of aortic valve involvement; 3) the persistence of heart block, despite at least one week of optimal antibiotics; and 4) the elimination of other potential causes of conduction abnormalities.