The concentration-dependent metabolism of 1-(14)C-labelled precursors of 22:5n-6 and 22:6n-3 was compared in rat testis cells. The amounts of [(14)C]22- and 24-carbon metabolites were measured by HPLC. The conversion of [1-(14)C]20:5n-3 to [3-(14)C]22:6n-3 was more efficient than that of [1-(14)C]20:4n-6 to [3-(14)C]22:5n-6. At low substrate concentration (4 microM) it was 3.4 times more efficient, reduced to 2.3 times at high substrate concentration (40 microM). The conversion of [1-(14)C]22:5n-3 to [1-(14)C]22:6n-3 was 1.7 times more efficient than that of [1-(14)C]22:4n-6 to [1-(14)C]22:5n-6 using a low, but almost equally efficient using a high substrate concentration. When unlabelled 20:5n-3 was added to a cell suspension incubated with [1-(14)C]20:4n-6 or unlabelled 22:5n-3 to a cell suspension incubated with [1-(14)C]22:4n-6, the unlabelled n-3 fatty acids strongly inhibited the conversion of [1-(14)C]20:4n-6 or [1-(14)C]22:4n-6 to [(14)C]22:5n-6. In the reciprocal experiment, unlabelled 20:4n-6 and 22:4n-6 only weakly inhibited the conversion of [1-(14)C]20:5n-3 and [1-(14)C]22:5n-3 to [(14)C]22:6n-3. The results indicate that if both n-6 and n-3 fatty acids are present, the n-3 fatty acids are preferred over the n-6 fatty acids in the elongation from 20- to 22- and from 22- to 24-carbon atom fatty acids. In vivo the demand for 22-carbon fatty acids for spermatogenesis in the rat may exceed the supply of n-3 precursors and thus facilitate the formation of 22:5n-6 from the more abundant n-6 precursors.