Repeated critical swimming performance trials (Ucrit) were performed on Atlantic salmon (Salmo salar) to test the null hypothesis that the source of dietary lipids (fish-based, poultry-based, and plant-based) does not influence exercise and recovery performance. Four diets were prepared by extensively replacing supplemental lipid from anchovy oil (AO; 100% AO at 150 g/kg) with cold pressed flaxseed oil (FO; 25% AO, 75% FO), sunflower oil (SO; 25% AO, 75% SO), or poultry fat (PF; 25% AO, 75% PF). These diets had equivalent protein and energy concentrations, but due to the different supplemental lipid sources, varied widely in their fatty acid composition. Fish fed AO had a significantly higher (P<0.05) first Ucrit (2.62Â±0.07 body lenght s-1) than those fed PF (2.22Â±0.12 body lenght s-1) that had low muscle ratios of n-3 highly unsaturated fatty acids (n-3 HUFA) to saturated fatty acids (SFA) and arachidonic acid (AA), and high levels of oleic acid. Fish in the FO and SO diet groups swam as well as AO-fed fish in both swimming trials. The performance of fish fed AO decreased significantly (P<0.05) during the second swimming trial (i.e. Ucrit2/Ucrit1=0.92Â±0.02). No significant differences occurred between diet groups for the second swim trial. There was a positive correlation between both n-3 HUFA/SFA and n-3 HUFA/AA ratios, and Ucrit1. A negative correlation was found between dietary AA and oleic acids, and Ucrit1. The present study suggests that low dietary n-3 HUFA/ SFA and n-3 HUFA/AA ratios may negatively affect swimming performance. The former possibly can be offset by increasing linoleic acid in the presence of nutritionally adequate n-3 HUFA (e.g. SO diet). Lipid supplements consisting largely of vegetable oils did not compromise fish cardiorespiratory physiology under the conditions of this study.