Pulmonary vascular tone was investigated by the construction of pulmonary arterial pressure (PAP)/cardiac output (Q) plots, and gas exchange, by the multiple inert gas elimination technique, in 24 anesthetized dogs before and after pulmonary embolization of autologous clots. Three PAP/Q plots were obtained by a manipulation of venous return at baseline and 60 min and 110 min after embolization. Before the third PAP/Q plot, the dogs were randomly allocated to one of the following iv treatments: 1) placebo (n = 6); 2) prostaglandin E1 (PGE1) 0.4 microgram.kg-1.min-1 (n = 6); 3) hydralazine 2 mg/kg (n = 6); and 4) nitroprusside 10 microgram.kg-1.min-1 (n = 6). These vasodilators decreased systemic arterial pressure by a mean of 44%. Ventilation-perfusion (VA/Q) distributions were determined at the same Q (2.4 +/- 0.1 l.min-1.m-2, mean +/- SE) of each PAP/Q plot. Embolization increased the intercept and the slope of the PAP/Q plots (P less than 0.001). Distributions of VA/Q were only moderately impaired, with an increased dispersion of both VA and Q and a shift of VA distributions to higher VA/Q. PaO2 changed from 208 +/- 5 to 172 +/- 8 mmHg (P less than 0.01) (fraction of inspired O2 was 0.4). None of the treatments had any effect on VA/Q distributions. Placebo and PGE1 had no effect on PAP/Q plots. Hydralazine and nitroprusside reduced the slope of the PAP/Q plots. Thus, in this canine model of acute pulmonary embolism: 1) VA/Q distributions were moderately impaired accounting for only slight hypoxemia, and 2) pulmonary hypertension was partially reversible by hydralazine and by nitroprusside without associated non-flow-dependent change in VA/Q distributions and arterial oxygenation.