An increase in atrial pressure has been shown to cause an increase in the concentration of atrial peptides (atriopeptin) in plasma. We therefore hypothesized that a reduction in atrial pressure would decrease the concentration of atriopeptin in plasma. In formulating this hypothesis we assumed that changes in the concentration of other circulating hormones or changes in cardiac nerve activity during hemorrhage would not affect the secretion of atriopeptin. To test the hypothesis, we bled sham-operated conscious dogs at a rate of 0.8 ml.kg-1.min-1 to decrease right and left atrial pressures. Hemorrhage was continued until a total of 30 ml of blood per kilogram body weight had been removed. Identical experiments were performed on conscious cardiac-denervated dogs. The concentration of plasma atriopeptin was decreased in each group of dogs after 10 ml of blood per kilogram of body weight had been removed, but the decrease achieved statistical significance only in the cardiac-denervated dogs. Further hemorrhage, however, produced no further decreases in circulating atriopeptin in either group even though atrial pressures continued to decline as more blood was removed. A comparison of the atriopeptin response to hemorrhage revealed no significant difference between the sham-operated and cardiac-denervated dogs, thus providing no evidence for a specific effect of cardiac nerves on atriopeptin secretion during hemorrhage. Our results demonstrate that the relationship between atrial pressure and plasma atriopeptin that has been observed repeatedly during atrial stretch is not evident during relatively slow, prolonged hemorrhage. There is, however, a small decline in circulating atriopeptin during the initial stage of hemorrhage that could be of biological significance.