Fatty acids of plasma and red cells of preterm babies, gestational age 28-33 weeks, weighing less than 2200 g were studied between birth and the expected date of delivery (EDD). Babies were fed either mothers' breast milk, or if they were unable, or chose not to breastfeed, randomly assigned to milk formula A or B. Milk B had 26% oleic acid, 0.5% docosahexaenoic acid (DHA) and 0.12% arachidonic (AA); A had 13% oleic acid, less than 0.05% AA and no DHA. The proportions of gamma linoleic and the other fatty acids, and the ratio of linoleic acid (LA)/alpha linolenic acid (ALA) were comparable in the two formulae. Diet and/or physiological adaptation seemed to be responsible for the significant changes in the proportions of certain fatty acids between birth and EDD. Incorporation of DKA in formula B increased its concentration in plasma and red cells, and reduced the conversion of LA to AA. With respect to maintenance of DHA status, milk B was superior to milk A but did not match breast milk. Conclusion Fortification of formula with DHA without concomitant incorporation of AA may precipitate AA insufficiency and may have developmental implications.