Our goal was to test directly whether the vasoconstrictor action of naloxone during hypovolemic hypotension is centrally mediated. In eight chronically instrumented rabbits, progressive central hypovolemia and fall in cardiac output (CO) were produced by gradually inflating a cuff on the thoracic vena cava. Central hypovolemia was then sustained for 8 min by holding CO constant. In the main experiment (n = 4), each rabbit was studied eight times over 4 experimental days. Saline or naloxone treatment commenced 10 min before progressive hypovolemia (early treatment) or 2 min after the onset of sustained hypovolemia (late treatment), given by intravenous infusion or into the fourth ventricle (V₄). With saline treatment, there was spontaneous recovery of systemic vasoconstriction and arterial pressure during sustained hypovolemia. Late treatment with naloxone (4 mg/kg iv; 4-37 µg/kg V₄) accelerated and exaggerated these changes. Thus, under conditions of constant CO and central blood volume, the vasodilatation of the decompensatory phase of acute hypovolemia is not sustained, and intravenous naloxone's vasoconstrictor action is via a brain stem mechanism.