Vascular resistance in canine limb skeletal muscle first increases and then decreases with prolonged arterial hypoxia, but whether neural sympathetic activity decreases with time is unknown. To assess the effectiveness of neurally mediated vasoconstrictor tone, we periodically cooled and rewarmed the sciatic nerve while nine anesthetized, paralyzed, pump-ventilated dogs were made hypoxic for 60 min by ventilation with 9.1% O2 in N2 (PaO2 = 24 +/- 2 mmHg). Before hypoxia, limb blood flow (QL) increased to a mean peak value of 111 ml.kg-1.min-1 with nerve cooling. With hypoxic hypoxia (HH), cardiac output increased but mean arterial pressure and limb blood flow remained the same. Nerve cooling at 15, 30, and 60 min of HH resulted in a pattern of progressively increasing mean peak QL values of 137, 151, and 160 ml.kg-1.min-1, respectively (P less than 0.05). Stimulation of the cut sciatic nerve at the end of the experiment established the maximum vasoconstriction that was possible and, thereby, the potential range that was available. The results showed that not only was neurally mediated vasoconstriction to skeletal muscle maintained throughout the hypoxic period, but that its intensity must have been increasing to overcome the local vasodilatory forces that were responsible for flow increasing even further with nerve cooling in prolonged hypoxia.