1. The effects of cirrhosis on mesenteric vascular reactivity were assessed in constantly perfused mesenteric arterial beds isolated from cirrhotic rats (carbon tetrachloride with phenobarbitone, n = 6), and from phenobarbitone-treated and untreated age-matched controls (n = 4,5). 2. At a constant flow rate of 5 ml min-1 there was no difference in basal perfusion pressure between the groups. Electrical field stimulation (EFS; 4-32 Hz, 90V, 1 ms, 30 s) of perivascular nerves caused frequency-dependent increases in perfusion pressure which were not different between the groups. Dose-dependent vasoconstrictor responses to exogenous noradrenaline (NA), methoxamine (an alpha 1-adrenoceptor agonist), adenosine 5'-triphosphate (ATP) and vasopressin were also similar between the groups. 3. The nitric oxide (NO) synthesis inhibitor NG-nitro-L-arginine methyl ester (L-NAME; 30 microM) augmented constrictor responses to NA, EFS, methoxamine and vasopressin in all groups, and as shown for EFS and NA, this was reversed by L-arginine (300 microM). However, the maximum constrictor responses of cirrhotic preparations in the presence of L-NAME were significantly lower than those of both groups of control animals at the highest frequency of EFS (32 Hz) and highest doses of NA (0.15 and 0.5 mumol) and, compared to phenobarbitone-treated controls, methoxamine (5 mumol). Responses to ATP were significantly augmented by L-NAME only in the cirrhotic group. 4. A step-wise increase in perfusate flow to 10, 15 and 20 ml min-1 produced a broadly similar increase in perfusion pressure within each group. At increased flow rates, cirrhotic preparations were hyporesponsive to NA (15 nmol) compared to the phenobarbitone-treated animals but not the untreated controls. Glibenclamide (5 microM) or L-NAME (30 microM) had no significant effect on the relationship between flow and perfusion pressure or on responses to NA at the different flow rates. 5. We conclude that sympathetic neurotransmission is unchanged in cirrhosis. Endogenous NO is important in modulation of constriction in both normal and cirrhotic states. Changes in NO may occur in cirrhosis, although the role of this in hyporesponsiveness of cirrhotic preparations to NA at higher flow rates and to the greater potentiation of ATP-mediated constriction in the presence of L-NAME, together with the impact of factors such as changes in calcium and potassium channels, is not entirely clear.