Adding docosahexaenoic acid (DHA) (22:6n-3) to formulas is more effective than increasing formula alpha-linolenic acid (18:3n-3) in maintaining blood phospholipid DHA levels similar to those in breast-fed infants. However, in long-term trials supplementary DHA given as marine oil reduces blood phospholipid arachidonic acid (AA) in preterm infants. This effect is not seen in short-term trials unless the total n-3 intake from marine oil exceeds 0.5% of the total fatty acids. In addition, there is considerable variability among individual preterm infants in blood phospholipid AA and DHA levels that is not dependent on diet. Within dietary treatments, a significant positive correlation between AA and DHA concentrations suggests that factor(s) other than marine oil supplementation affect both AA and DHA status. Docosahexaenoic acid and AA concentrations in plasma phospholipids are significantly correlated with DHA and AA concentrations in red blood cell phospholipids, suggesting that the observed individual differences in DHA and AA within groups represent true differences in fatty acid status. Preterm infants appear to be vulnerable to a poor status of both DHA and AA; further feeding trials are needed to identify the optimal balance of fatty acids for feeding these infants.