According to the revised pathway for 22:6(n - 3) biosynthesis in liver (Voss et al. (1991) J. Biol. Chem. 266, 19995-20000) both 18:3(n - 3) and 24:5(n - 3) serve as substrates for desaturation at position-6. The present study was undertaken to determine whether microsomes contain chain-length-specific 6-desaturases. Addition of [1-14C]20:3(n - 6), a substrate for desaturation at position-5, did not depress desaturation of either [1-14C]18:3(n - 3) or [3-14C]24:5(n - 3). An unexplained observation was that both 18:3(n - 3) and 24:5(n - 3) inhibited the metabolism of 20:3(n - 6) to 20:4(n - 6). When an enzyme-saturating level of [3-14C]24:5(n - 3) was now incubated alone and with 40, 80 and 120 nmol of [1-14C]18:3(n - 3), the production of 24:6(n - 3) was inhibited by 43, 67 and 81%. Conversely, when [1-14C]18:3(n - 3) was incubated with 40, 80 or 120 nmol of [3-14C]24:5(n - 3), the synthesis of 18:4(n - 3) was inhibited by only 15, 20 and 27%. These and other competitive studies showed that there was always preferential desaturation of 18:3(n - 3) rather than 24:5(n - 3). In addition, competitive studies between 18:2(n - 6) and 18:3(n - 3), as well as with 24:4(n - 6) and 24:5(n - 3) showed that there was always preferential desaturation of the (n - 3) acid. Although our results are consistent with a single 6-desaturase, further studies, including the isolation of the 6-desaturases(s), is obviously required to determine whether multiple forms of the 6-desaturase exist.