The synthesis of 4,7,10,13,16,19-docosahexaenoic acid (22:6(n-3)) requires that when 6,9,12,15,18,21-tetracosahexaenoic acid (24:6(n-3)) is produced in the endoplasmic reticulum, it preferentially moves to peroxisomes for one cycle of beta-oxidation rather than serving as a substrate for membrane lipid synthesis. Both 24:6(n-3) and its precursor, 9,12,15,18,21-tetracosapentaenoic acid (24:5(n-3)), were poor substrates for acylation into 1-acyl-sn-glycero-3-phosphocholine (1-acyl-GPC) by rat liver microsomes. When peroxisomes were incubated with 1-14C- or 3-14C-labeled 7,10,13,16,19-docosapentaenoic acid (22:5(n-3)), [1-14C]22:6(n-3), [3-14C]24:5(n-3), or [3-14C]24:6(n-3), only small amounts of acid-soluble radioactivity were produced when double bond removal at positions 4 and 5 was required. When microsomes and 1-acyl-GPC were included in incubations, the preferred metabolic fate of acids, with their first double bond at either positions 4 or 5, was to move out of peroxisomes for esterification into the acceptor rather than serving as substrates for continued beta-oxidation. When [1-14C]22:6(n-3) or [3-14C]24:6(n-3) was incubated with peroxisomes, 2-trans-4,7,10,13,16,19-22:7 accumulated. The first cycle of 20:5(n-3) beta-oxidation proceeds through 2-trans-4,8,11,14,17-20:6 and thus requires both Delta3,5,Delta2, 4-dienoyl-CoA isomerase and 2,4-dienoyl-CoA reductase. The accumulation of the substrate for 2,4-dienoyl-CoA reductase, as generated from 22:6(n-3), but not from 20:5(n-3), suggests that this enzyme distinguishes between subtle structural differences. When 22:6(n-3) is produced from 24:6(n-3), its continued degradation is impaired because of low 2,4-dienoyl-CoA reductase activity. This slow reaction rate likely contributes to the transport of 22:6(n-3) out of peroxisomes for rapid acylation into 1-acyl-GPC by microsomes.