We tested the hypothesis that arterial baroreceptor discharge remains constant during nonhypotensive hemorrhage. Aortic baroreceptor activity was recorded from anesthetized dogs during slow, continuous hemorrhage until 24 ml blood/kg body wt had been removed. The relationship between simultaneously recorded hemodynamic variables and single-unit aortic baroreceptor activity was evaluated by a unique computer program that sampled all variables 500 times/s and analyzed data from each consecutive cardiac cycle throughout the entire experiment. In three hemorrhages performed on three dogs, aortic blood pressure did not decrease during the experiment (nonhypotensive hemorrhage), but aortic nerve activity decreased progressively when data from individual cardiac cycles having identical mean aortic pressures were compared. Reduced baroreceptor activity correlated closely with progressive reductions in pulse pressure. In other hemorrhages, mean aortic pressure decreased by a variable amount; decreases in recorded single-unit baroreceptor activity occurred during each of these hemorrhages also. In summary, hemorrhage consistently caused decreases in the activity of aortic baroreceptors regardless of whether mean aortic pressure decreased or remained constant during hemorrhage. We conclude that experimental techniques employing nonhypotensive hemorrhage do not provide an effective method for selectively studying reflexes from cardiopulmonary receptors in the absence of changes in arterial baroreceptor input.