The metabolism of individual dietary n-3 fatty acids was studied in n-3 fatty acid-deficient newly hatched chicks. Laying hens were fed the n-6 fatty acid, ethyl linoleate, as the only source of polyunsaturated fat. Chicks were then fed the n-3-deficient hens' diet, or one of three other diets supplemented with the ethyl ester of 18:3 n-3, 20:5 n-3 [eicosapentaenoic acid (EPA)], or 22:6 n-3 [docosahexaenoic acid (DHA)] at 0.44% of calories. At the end of 0, 1, 2, and 3 wk, the fatty acid composition of the brain, retina, liver, and serum was determined. Dietary EPA and DHA were equally effective at raising levels of DHA in the brain and retina. Dietary 18:3 was relatively ineffective in restoring brain and retina DHA. In the n-3-deficient chicks fed EPA or DHA, levels of DHA recovered to control values in both the brain and retina by 3 wk. Very little EPA accumulated in the brain or retina of chicks fed EPA. Hepatic synthesis of DHA from EPA appeared low, suggesting that the brain and retina synthesized the DHA that accumulated rapidly in these tissues after the feeding of EPA. The delta-4-desaturase enzyme was apparently very active, then, in the brain and retina. Retroconversion of dietary 22:6 to 22:5 and 20:5 was evident in the serum, liver, and retina but not in the brain. Thus, it was possible to study the relative metabolism and especially the interconversion of n-3 fatty acids in a environment uncomplicated by existing stores of these essential fatty acids. This study would suggest that 18:3 as the sole source of n-3 fatty acids in the diets of animals, including the human infant, may not be adequate for the biochemical development of the brain and retina and that dietary DHA is the preferred fatty acid of the n-3 series.