1. The aims of this study in the ovine fetus were to (1) characterise continuous changes in umbilical blood flow and vascular conductance during acute hypoxaemia and (2) determine the effects of nitric oxide blockade on umbilical blood flow and vascular conductance during normoxic and hypoxaemic conditions using a novel in vivo 'nitric oxide clamp'. 2. Under 1-2% halothane anaesthesia, seven ovine fetuses were instrumented between 118 and 125 days of gestation (term is ca 145 days) with vascular and amniotic catheters and a flow probe around an umbilical artery. At least 5 days after surgery, all fetuses were subjected to a 3 h protocol: 1 h of normoxia, 1 h of hypoxaemia and 1 h of recovery during fetal I.V. infusion with saline or, 1-2 days later, during combined fetal treatment with the nitric oxide (NO) inhibitor N (G)-nitro-L-arginine methyl ester (L-NAME, 100 mg x kg(-1)) and the NO donor sodium nitroprusside (NP, 5.1 +/- 2.0 microg x kg(-1) x min(-1), the 'nitric oxide clamp'). Following the end of the 3 h experimental protocol, the infusion of NP was withdrawn to unmask any persisting effects of fetal treatment with L-NAME alone. 3. During acute hypoxaemia, the reduction in arterial partial pressure of O2 (Pa,O2) was similar in fetuses infused with saline or treated with the nitric oxide clamp. In all fetuses, acute hypoxaemia led to a progressive increase in mean arterial blood pressure and a fall in heart rate. In saline-infused fetuses, acute hypoxaemia led to a rapid, but transient, decrement in umbilical vascular conductance. Thereafter, umbilical vascular conductance was maintained and a significant increase in umbilical blood flow occurred, which remained elevated until the end of the hypoxaemic challenge. In contrast, while the initial decrement in umbilical vascular conductance was prevented in fetuses treated with the nitric oxide clamp, the increase in umbilical blood flow during hypoxaemia was similar to that in fetuses infused with saline. After the 1 h recovery period of the acute hypoxaemia protocol, withdrawal of the sodium nitroprusside infusion from fetuses undergoing the nitric oxide clamp led to a significant, but transient, hypertension and a sustained umbilical vasoconstriction. 4. In conclusion, the data reported in this study of unanaesthetised fetal sheep (1) show that minute-by-minute analyses of haemodynamic changes in the umbilical vascular bed reveal an initial decrease in umbilical vascular conductance at the onset of hypoxaemia followed by a sustained increase in umbilical blood flow for the duration of the hypoxaemic challenge, (2) confirm that the increase in umbilical blood flow after 15 min hypoxaemia is predominantly pressure driven, and (3) demonstrate that nitric oxide plays a major role in the maintenance of umbilical blood flow under basal, but not under acute hypoxaemic, conditions.