The purpose of this study was to test the assumption that the cause for postoperative atrial flutter/fibrillation after cardiopulmonary bypass is inadequate atrial myocardial protection. Dogs were subjected to cardioplegic arrest for 60 minutes without augmented atrial hypothermia (seven dogs, control group) or augmented atrial hypothermia with topical atrial cooling (seven dogs, study group). Twenty-five electrodes (15 on the right atrium and 10 on the left atrium) were fixed on the atria to measure effective refractory period and conduction time. Data were taken before bypass, immediately after bypass, and 2 hours after bypass. During cardioplegic arrest the mean temperatures measured in the atria were significantly lower (p < 0.001) in the study group (13.5° ± 7.0° C) than in the control group (23.7° ± 3.2° C). There was no significant change in the mean effective refractory period after bypass in the control or study groups or in the prevalence of inducibility of atrial flutter/fibrillation by extrastimulation (3/7 dogs in the control group and 2/7 in the study group). During right atrial pacing, total conduction times were significantly longer (p < 0.025 at cycle lengths of 300 and 350 msec) in the control group (74 ± 5 msec and 75 ± 7 msec, respectively) than in the study group (65 ± 9 msec and 64 ± 8 msec, respectively) immediately after bypass. Two hours after bypass, however, there were no significant differences under the same conditions between the two groups. There were no significant differences in conduction during left atrial pacing after bypass. Comparing those atria that were inducible with those not inducible demonstrated a significantly increased dispersion of effective refractory period (90 ± 23 msec versus 74 ± 18 msec, p < 0.05) and increased conduction time in the inducible group. We concluded that augmented atrial hypothermia during cardioplegic arrest had no effect on the inducibility of fibrillation, had no effect on repolarization, and had only a small effect on conduction, which resolved within 2 hours after bypass. However, the study demonstrates that when the atria are inducible the substrates are an increased dispersion of refractoriness and a prolongation of conduction time.