We investigated whether the reflex responses to stimulation of pulmonary arterial baroreceptors were altered by intrathoracic pressure changes similar to those encountered during normal breathing. Dogs were anaesthetized with -chloralose, a cardiopulmonary bypass was established, and the pulmonary trunk and its main branches as far as the first lobar arteries were vascularly isolated and perfused with venous blood. The chest was closed following connection to the perfusion circuit and pressures distending the aortic arch, carotid sinus and coronary artery baroreceptors were controlled. Changes in the descending aortic (systemic) perfusion pressure (SPP; flow constant) were used to assess changes in systemic vascular resistance. Values of SPP were plotted against mean pulmonary arterial pressure (PAP) and sigmoid functions applied. From these curves we derived the threshold pressures (corresponding to 5% of the overall response of SPP), the maximum slopes (equivalent to peak gain) and the corresponding PAP (equivalent to ‘set point’). Stimulus–response curves were compared between data obtained with intrathoracic pressure at atmospheric and with a phasic intrathoracic pressure ranging from atmospheric to around -10 mmHg (18 cycles min-1). Results were obtained from seven dogs and are given as means ±S.E.M. Compared to the values obtained when intrathoracic pressure was at atmospheric, the phasic intrathoracic pressure decreased the pulmonary arterial threshold pressure in five dogs; average change from 28.4 ± 5.9 to 19.3 ± 5.9 mmHg (P > 0.05). The inflexion pressure was significantly reduced from 37.8 ± 4.8 to 27.4 ± 4.0 mmHg (P < 0.03), but the slopes of the curves were not consistently changed. These results have shown that a phasic intrathoracic pressure, which simulates respiratory oscillations, displaces the stimulus–response curve of the pulmonary arterial baroreceptors to lower pressures so that it lies within a physiological range of pressures.