The fatty acid composition of erythrocyte membranes was investigated in 21 healthy men after 6 wk of varying intakes of eicosapentaenoic acid (EPA, 20:5n-3) and docosahexaenoic acid (DHA, 22:6n-3). In one experiment, 12 subjects were fed three diets in a 3 x 3 crossover design: an essentially fish-free control diet, a fish diet (0.15 g EPA/d, 0.41 g DHA/d) and the same fish-based diet supplemented with 5 g/d fish oil (Fish + Oil: 0.99 g EPA/d, 0.99 g DHA/d). A 6 wk wash-out period was allowed between each diet. In another experiment, 11 subjects were supplemented with 5 g/d fish oil alone for 6 wk (0.84 g EPA/d, 0.48 g DHA/d). After fish or fish oil feeding, the percent proportion of EPA and DHA in the erythrocyte membranes rose at the expense of linoleic and arachidonic acids. After 6 wk on the fish-based diets, EPA incorporation approached saturation, with the incremental increases being proportional to the amounts supplied by the diets. In contrast, parallel increases were observed for erythrocyte DHA even though the Fish + Oil diet was supplying twice as much DHA as the fish alone diet. These observations imply different metabolic rates for EPA and DHA and their importance is discussed in terms of the value of erythrocyte EPA versus DHA as markers for fish and fish oil consumption.