1. Noradrenaline sensitivity and acetylcholine-induced relaxation were investigated in mesenteric resistance arteries of control and streptozotocin-induced diabetic rats. 2. The diabetic rats demonstrated enhanced vascular sensitivity to noradrenaline compared with age-matched controls (pEC50 5.99 +/- 0.06 for diabetic rats, n = 25, versus 5.82 +/- 0.03 for controls, n = 45, P < 0.05). 3. Significant impairment of acetylcholine-induced relaxation was observed in arteries from the diabetic animals compared with controls (pEC50 6.81 +/- 0.17 for diabetic rats, n = 21, versus 7.54 +/- 0.17 for controls, n = 45, P < 0.001). 4. The difference between acetylcholine-induced relaxation in diabetic and control arteries remained in the presence of 10 microM indomethacin (pEC50 6.41 +/- 0.11 for diabetic rats, n = 16, versus 7.59 +/- 0.08 for controls, n = 20, P < 0.001). 5. The nitric oxide synthase inhibitor, NG-monomethyl-L-arginine (L-NMMA, 1 mM) produced profound inhibition of acetylcholine-induced relaxation in diabetic arteries but partial inhibition in controls. The incomplete inhibition of acetylcholine-induced relaxation by L-NMMA in the control arteries was the result of ineffective inhibition of nitric oxide synthase since an alternative inhibitor, NG-nitro-L-arginine methyl ester (L-NAME, 0.1 mM), led to similar inhibition to that seen in the diabetic arteries with L-NMMA. The endothelium-derived relaxing factor (EDRF)-mediated component of acetylcholine-induced relaxation determined by use of the nitric oxide synthase inhibitors was, therefore, apparently reduced in diabetic rats compared with control animals.6. In further experiments L-NAME was found to enhance the response to noradrenaline in control rats but not in diabetic animals, suggesting that the abnormal response to noradrenaline in the diabetic animals was also due to reduced EDRF release.7. Nitroprusside-induced relaxation (endothelium-independent) was similar in arteries from control anddiabetic rats (pEC5o 7.61 +/- 0.13 for diabetic arteries, n = 18, versus 7.68 +/- 0.15 in the controls, n = 20, P not significant).8. These results suggest that endothelial function is abnormal in mesenteric resistance arteries of streptozotocin-induced diabetic rats and that this is predominantly due to reduced EDRF release.