The circulatory responses to different intravenous doses of naloxone were studied in conscious rabbits before and after haemorrhage, under different conditions including prior ganglion blockade. Unless there had been blood loss, naloxone elicited no pressor response, even in high dose. After bleeding so that arterial pressure fell to 40 mmHg, the dose-response relationship for naloxone had two components. Over a low-dose range (threshold 0.3 mg/kg) naloxone had a modest pressor effect but did not affect heart rate. Over a much higher dose range (threshold 0.6 mg/kg) naloxone caused a marked rise in arterial pressure and a profound bradycardia. The highest dose of naloxone examined (25 mg/kg) caused a rise in arterial pressure of 70 mmHg and a reduction in heart rate of 160 beats/min. The pressor and bradycardic effects of naloxone were the same whether post-haemorrhagic hypotension lasted 5, 10, 20 or 30 min. The responses to naloxone in low or high dose depended much more closely on the volume of blood removed than on the level to which arterial pressure fell. Even after non-hypotensive haemorrhage a high dose of naloxone had marked pressor and bradycardic effects. Ganglion blockade prior to haemorrhage abolished the pressor response to a low, but not to a high, dose of naloxone. It was concluded that prolonged and severe hypotension are not necessary to 'prime' the cardiovascular system to respond to naloxone after haemorrhage. In a high dose its pressor effects appear to be mediated post-ganglionically, but in a low dose it may act within the central nervous system.