1. The vasodilator effects of hydralazine in vitro, using the Krebs' perfused human placental lobule was studied. Single placental lobules were bilaterally perfused (maternal and fetal sides 5 mL/min each, 95% O2, 5% CO2, 37 degrees C) and changes in fetal arterial pressure (FAP) and venous outflow (VO) were recorded. 2. Submaximal vasoconstriction was induced by KCl (20-50 mmol/L), which increased basal FAP from 22.8 +/- 1.7 to 91.3 +/- 3.9 mmHg (n = 9, P < 0.001), and decreased VO from 4.1 +/- 0.6 to 0.2 +/- 0.1 mL/min (n = 6, P < 0.01). 3. Hydralazine caused vasodilatation (IC50 1.9 mmol/L, n = 9) and increased VO in the presence of KCl-induced vasoconstriction. 4. Infusion of N omega-nitro-L-arginine (100 mumol/L) to block nitric oxide synthase caused the basal FAP to increase from 30.9 +/- 5.9 to 47.4 +/- 6.7 (n = 6, P < 0.05) and significantly potentiated hydralazine-induced vasodilatation (n = 7, P < 0.05). 5. The soluble guanylate cyclase inhibitor LY 83583 (6-anilino-5,8-quinolinedione) (1 mumol/L) significantly antagonized the vasodilatation produced by hydralazine (n = 5, P < 0.05). 6. Thus, Hydralazine appears to activate guanylate cyclase, leading to increased cyclic GMP in fetal arterial vascular smooth muscle to cause vasorelaxation. No evidence was obtained to suggest that hydralazine exerted its action by either releasing nitric oxide from endothelial cells in the placenta or acting as a nitric oxide donor.