Our question was whether inhibition of nitric oxide [endothelium-derived relaxing factor (EDRF)/NO] production in an in situ vascularly isolated but innervated canine hindlimb would prevent hypoxic vasodilation or interfere with O2 extraction during ischemic (IH) or hypoxic hypoxia (HH). After a control period, we gave NG-nitro-L-arginine methyl ester (L-NAME, 20 mg/kg i.v.) to two of four groups of six dogs before a 30-min period of IH or HH. In IH, arterial inflow from a pump-membrane oxygenator system was lowered from 65 to 35 ml.min-1.kg-1 with PO2 maintained at approximately 110 Torr. In HH, PO2 was lowered from 107 to 28 Torr with flow at 78 ml.min-1.kg-1. Total O2 delivery was lowered to approximately 5 ml.min-1.kg-1 in all groups during hypoxia. Hindlimb vascular resistance (LVR) increased from 1.11 +/- 0.09 to 2.21 +/- 0.25 peripheral resistance units (PRU; P < 0.05) after L-NAME infusion and hindlimb O2 uptake increased from 3.9 +/- 0.2 to 4.5 +/- 0.3 ml.min-1.kg-1 (P < 0.05). In controls, LVR decreased from 1.10 +/- 0.06 to 0.63 +/- 0.04 PRU with HH (P < 0.05) and from 1.03 +/- 0.06 to 0.82 +/- 0.02 PRU (P = NS) with IH. In L-NAME-treated dogs, LVR decreased from 2.38 +/- 0.37 to 1.07 +/- 0.13 PRU with HH (P < 0.05) and from 2.04 +/- 0.29 to 1.41 +/- 0.13 PRU (P = NS) with IH. There were no differences in O2 extraction ratio (0.72) or in O2 uptake between groups during hypoxia.(ABSTRACT TRUNCATED AT 250 WORDS)