Hypercapnia has been shown in animal experiments to induce pulmonary hypertension. This study measured the sensitivity and time course of the human pulmonary vascular response to sustained (4 h) hypercapnia and hypocapnia. Twelve volunteers undertook three protocols: 1) 4-h euoxic (end-tidal Po(2) = 100 Torr) hypercapnia (end-tidal Pco(2) was 10 Torr above normal), followed by 2 h of recovery with euoxic eucapnia; 2) 4-h euoxic hypocapnia (end-tidal Pco(2) was 10 Torr below normal) followed by 2 h of recovery; and 3) 6-h air breathing (control). Pulmonary vascular resistance was assessed at 0.5- to 1-h intervals by using Doppler echocardiography via the maximum tricuspid pressure gradient during systole. Results show progressive changes in pressure gradient over 1-2 h after the onset or offset of the stimuli, and sensitivities of 0.6 to 1 Torr change in pressure gradient per Torr change in end-tidal Pco(2). The human pulmonary circulatory response to changes in Pco(2) has a slower time course and greater sensitivity than is commonly assumed. Vascular tone in the normal pulmonary circulation is substantial.