A factorial, two-way, experimental design was used for this 10-week nutritional trial, aiming to elucidate the interactive effects of decreasing dietary protein:lipid level and substitution of fish oil (FO) with rapeseed oil (RO) on tissue fatty acid (FA) composition and metabolism of large Atlantic salmon (Salmo salar L.) reared at high water temperatures (sub-optimal, summer temperatures: 11·6°C). The six experimental diets were isoenergetic and formulated to include either FO or RO (60 % of the added oil) at three dietary protein:lipid levels, specifically (1) 350 g/kg protein and 350 g/kg lipid, (2) 330 g/kg protein and 360 g/kg lipid, (3) 290 g/kg protein and 380 g/kg lipid. Final weight, specific growth rate and thermal growth coefficient were positively affected by the dietary RO inclusion at the expense of FO, while no significant effects were seen on growth due to the decreasing protein level. The oil source had a significant effect on muscle and liver FA composition. However, the changes in muscle and liver FA indicate selective utilisation or retention of individual FA and moderate reductions in tissue EPA and DHA. Pyloric caeca phospholipid FA composition was significantly affected by the two factors and, in some cases, significant interactions were also revealed. Liver and red muscle β-oxidation capacities were significantly increased due to RO inclusion, while an interactive effect of protein level and oil source was shown for white muscle β-oxidation capacity. The results could explain, at least partially, the better performance that was shown for the RO groups and the enhanced protein-sparing effect.