Intraatrial reentrant tachycardia, or atrial flutter, is a common postoperative problem after Fontan repair, which involves an atriopulmonary connection. A modification of Fontan repair, total cavopulmonary connection, minimizes the portion of the right atrium exposed to stretch and hypertension; however, atrial flutter continues to occur after this procedure. We postulated that the intraatrial lateral tunnel suture line of total cavopulmonary connection, in the absence of physiologic alterations such as atrial hypertension or stretch, provides the necessary electrophysiologic substrate for atrial flutter. The purpose of this study was to produce a canine model of total cavopulmonary connection (1) to establish that the intraatrial suture line alone is sufficient to permit sustained atrial flutter and (2) to characterize the pathways of resulting reentrant arrhythmias. After induction of general anesthesia, 25 to 30 kg dogs (n = 17) underwent median sternotomy, cradling of the pericardium, and placement of a pacing electrode on the right atrial appendage. Normothermic cardiopulmonary bypass was initiated. The total cavopulmonary connection suture line was placed through a standard right atriotomy, simulating construction of the lateral tunnel. After closure of the atriotomy, 253 point unipolar atrial endocardial form-fitting electrodes were placed through bilateral ventriculotomies. By means of atrial burst pacing and programmed extrastimulation, induction of atrial flutter was attempted. If atrial flutter could not be induced, isoproterenol was infused and the stimulation protocol was repeated. After induction of atrial flutter, mapping of the activation sequence was performed. Before suture line placement, no dog had inducible atrial flutter. After placement of the suture line, sustained atrial flutter was reproducibly induced in every dog, although isoproterenol was required for this in three (17.6%). The mean flutter cycle length was 177 ± 30 msec. In each case, the atrial flutter circuit was limited to the right atrium, with the left atrium being passively activated. The atrial flutter circuit was dependent on a corridor of myocardium that resulted from conduction block on the free wall, created by the lateral margin of the total cavopulmonary connection. In no case was the atriotomy integral to the atrial flutter circuit. This study establishes that the total cavopulmonary connection baffle suture line alone, without alteration in circulatory physiology, creates a sufficient anatomic substrate for atrial flutter in a short-term canine model. Delineation of the anatomic boundaries of the reentrant circuit raises the possibility of targeting areas within the circuit that could be modified, potentially reducing the incidence of postoperative atrial flutter after total cavopulmonary connection.