The objective of this experiment was to test the effect of total or partial substitution of dietary fish oil (FO) by linseed oil (LO) in Atlantic salmon feeding on performance, liver and muscle fatty acid composition, selected lipogenic and lipolytic enzyme activities, and flesh oxidative stability. For 12 wk, fish (220 +/- 12 g of initial BW) were fed five experimental diets in which the FO was serially replaced by 25, 50, 75, and 100% LO. Total FO replacement by LO did not (P = 0.20) affect fish final weight, biometric indices, or i.m. fat contents. Liver and muscle neutral lipid (NL) composition responded to dietary treatments in different ways. Whereas the sum of n-3 PUFA in muscle followed a linear and quadratic pattern with increasing levels of LO, a linear (P = 0.005) effect was observed in the liver NL fraction. Total n-3 and n-6 PUFA contents in the polar lipid fraction (PL) were unaffected (P = 0.356) by dietary input of LO in muscle. Activity of liver glucose-6-P-dehydrogenase (G6PD) was greater with increasing levels of LO (P = 0.004). A time effect (P < 0.001) was observed in the concentration of lipid peroxidation products, expressed as thiobarbituric acid reactive substances, in fish flesh stored under refrigeration for 9 d; however, the progressive inclusion of LO in the feed did not affect (P = 0.125) flesh oxidation stability. In summary, LO can totally replace FO in Atlantic salmon feed without affecting growth performance and muscle susceptibility to lipid oxidation. Fatty acid metabolism in the liver was affected by LO, promoting G6PD activity and eicosatetraenoic acid accumulation; however, a 100% LO replacement decreased (P < 0.001) concentrations of eicosapentaenoic and docosahexaenoic acids in salmon muscle.