The aim of this study was to test our hypothesis that both phasic cardiac vagal activity and tonic pulmonary vagal activity, estimated as respiratory sinus arrhythmia (RSA) and anatomical dead space volume, respectively, contribute to improve the efficiency of pulmonary gas exchange in humans. We examined the effect of blocking vagal nerve activity with atropine on pulmonary gas exchange. Ten healthy volunteers inhaled hypoxic gas with constant tidal volume and respiratory frequency through a respiratory circuit with a respiratory analyser. Arterial partial pressure of O(2) (P(aO(2))) and arterial oxygen saturation (S(pO(2))) were measured, and alveolar-to-arterial P(O(2)) difference (D(A-aO(2))) was calculated. Anatomical dead space (V(D,an)), alveolar dead space (V(D,alv)) and the ratio of physiological dead space to tidal volume (V(D,phys)/V(T)) were measured. Electrocardiogram was recorded, and the amplitude of R-R interval variability in the high-frequency component (RRIHF) was utilized as an index of RSA magnitude. These parameters of pulmonary function were measured before and after administration of atropine (0.02 mg kg(-1)). Decreased RRIHF (P < 0.01) was accompanied by decreases in P(aO(2)) and S(pO(2)) (P < 0.05 and P < 0.01, respectively) and an increase in D(A-aO(2)) (P < 0.05). Anatomical dead space, V(D,alv) and V(D,phys)/V(T) increased (P < 0.01, P < 0.05 and P < 0.01, respectively) after atropine administration. The blockade of the vagal nerve with atropine resulted in an increase in V(D,an) and V(D,alv) and a deterioration of pulmonary oxygenation, accompanied by attenuation of RSA. Our findings suggest that both phasic cardiac and tonic pulmonary vagal nerve activity contribute to improve the efficiency of pulmonary gas exchange in hypoxic conscious humans.