Study objective - The mechanical coupling between the ventricles occurs directly through the myocardium (ventricular-ventricular coupling) and indirectly through the pericardium (ventricular-pericardial-ventricular coupling). We postulated that the magnitude of ventricular-pericardial-ventricular coupling would increase at high pericardial pressures, while ventricular-ventricular coupling would be unaltered.Design - Canine hearts were removed and placed in cold cardioplegic solution. Balloons were inserted into each ventricle and the left and right ventricular pressure (dP1, dPr) and volume (dV1, dVr) changes caused by increasing the pressure and volume of the other ventricle and by increasing pericardial pressure (dPp) were measured.Experimental materials - Hearts from 10 random source dogs, weight 12.5-18 kg, were used.Measurement and main results - At control pericardial pressure levels, the magnitude of the pericardial ventricular interactions was greater than the ventricular-ventricular interactions: dP1/dPp, was significantly greater than dP1/dPr, at 0.71 (SEM 0.04), n = 6, v 0.18(0.03), p<0.01, and dVl/dPp, was significantly greater than dVl/dPr, at -0.83(0.09) v -0.24(0.06), p<0.05. Raising the pericardial pressure increased the mechanical coupling between the ventricles: dPl/dPr∼, dVl/dPr∼, dPr/dPl∼, and dVr/dPl∼ increased significantly (p<0.05) by 0.48(0.03), 0.67(0.13), 0.38(0.05), and 0.61(0.09) respectively. This increased coupling occurred through pericardial pressure changes. If pericardial pressure was maintained constant, the coupling between the ventricles was unaltered. This same pattern was observed in four in situ experiments. For these experiments, at the raised pericardial pressure levels, dPl/dPr increased, from 0.51(0.03) to 0.79(0.01), p<0.05, if pericardial pressure was allowed to vary, but was unaltered with a constant pericardial pressure, at 0.42(0.03) v 0.44(0.04), p>0.5.Conclusions - Ventricular interdependence was increased with raised pericardial pressure and this increased coupling was due primarily to an increased ventricular-pericardial-ventricular coupling. This increased coupling may help to explain the paradoxical pulse observed in cardiac tamponade.