The effects of 3 months streptozotocin-induced diabetes mellitus on contraction and relaxation of aorta were examined in vitro. A further diabetic group was treated with a novel sulphonylnitromethane-based aldose reductase inhibitor for 3 months following diabetes induction. Diabetes resulted in reduced maximal tension production, particularly for responses to phenylephrine (p < 0.001) and serotonin (p < 0.001). However, with aldose reductase inhibitor treatment, responses were in the non-diabetic range. The ratio of maximum contractions to noradrenaline and phenylephrine were 28% elevated by diabetes (p < 0.01), which may suggest increased alpha2-adrenoreceptor-mediated responses. Endothelium-independent relaxation to glyceryl trinitrate was unaffected by diabetes or treatment. By contrast, there were 38% deficits in endothelium-dependent relaxation to acetylcholine (p < 0.001) and Ca2+ ionophore A23187 (p < 0.001) with diabetes which were prevented by aldose reductase inhibitor treatment (p < 0.001). A 121% shift in the concentration giving a 50% maximum effect for acetylcholine towards lower sensitivity with diabetes (p < 0.001) was also largely corrected by treatment (p < 0.001). A non-diabetic group treated with aldose reductase inhibitor showed a 30% decrease in the 50% effective concentration for acetylcholine (p < 0.05). A 15% deficit in maximum relaxation to the ATP-sensitive K+ channel opener cromakalim for the diabetic group (p < 0.001) was prevented by aldose reductase inhibitor treatment (p < 0.01). We conclude that there are polyol pathway related abnormalities for contraction, some aspects of endothelium-independent relaxation, but particularly for endothelium-dependent relaxation in aorta from chronic streptozotocin-diabetic rats. If found in the appropriate circulatory beds, these could potentially contribute to the putative vascular basis of some of the complications of diabetes. Their amelioration could account for many of the beneficial effects of aldose reductase inhibitors.