1. To examine possible mechanisms of antihypertensive effects of feeding fish oil rich in n-3 fatty acids, we have studied vascular reactivity of aortic rings and perfused mesenteric resistance vessels of spontaneously hypertensive rats (SHR) given such a diet. 2. In two experiments, rats were fed a semi-synthetic diet containing either 'fish oil' (10 and 20% by weight) or hydrogenated coconut oil (control) (10 and 20%) for 4 weeks. 3. Blood pressure rose significantly less in the fish oil group than in controls in both experiments. 4. Aortic rings from control rats showed endothelium-dependent relaxations to low concentrations of acetylcholine (ACh) but relaxed less at higher concentrations. In contrast, rings from the fish oil group had relaxations which increased through the range of concentrations used. Indomethacin (10 microM) also increased the relaxation responses seen in rings from control rats, suggesting that fish oil inhibits a contractile cyclo-oxygenase product. This contractile substance may be thromboxane A2 (TxA2) or its endoperoxide precursor, prostaglandin H2 (PGH2) as aortic incubates and serum levels of TxB2 (the stable product of TxA2) were greatly reduced in fish oil-fed rats, and the decrease of relaxant responses to high concentrations of ACh were also blocked by a TxA2/PGH2 receptor blocker (SQ 29548). 5. In contrast to aortic rings, perfused preconstricted mesenteric resistance vessels of control rats relaxed to ACh in a similar fashion to tissues from fish oil-fed rats. However, in this preparation, fish oil feeding enhanced relaxations to sodium nitroprusside (SNP) and contractile responses to noradrenaline were less than controls. After removal of endothelium with 0.05% saponin, contractile responses to noradrenaline increased in both groups but responses from fish oil-treated rats were still attenuated. This suggests that fish oil feeding alters reactivity of mesenteric resistance vessels at the level of the smooth muscle. 6. The results indicate that fish oil feeding may reduce blood pressure by decreasing vascular smooth muscle reactivity to noradrenaline in resistance vessels. The effect may be enhanced by inhibition of an endothelium-derived cyclo-oxygenase product, such as TxA2 or PGH2 in conduit vessels.