Between May 1993, and September 1994, a randomized, blinded clinical trial was conducted to evaluate measures of growth and body composition in 63 (32 males; 31 females) healthy, low‐birth‐weight infants (940–2250 g) who were randomly assigned to an infant formula with docosahexaenoic acid (22:6n3, DHA, 0.2 wt%) from fish oil or to a control formula. A preterm formula with or without DHA was fed beginning at 7–10 days prior to hospital discharge through 43 weeks postmenstrual age (PMA). Then, from 43‐59 weeks PMA, infants were fed a term infant formula with or without a corresponding amount of DHA. Growth (weight, length, head circumference), regional body fatness (triceps, subscapular, suprailiac skinfold thicknesses), circumferences (arm, abdominal, chest), and estimates of body composition determined by total body electrical conductivity (TOBEC) (fat‐free mass [FFM]) were evaluated. Growth was slower in males fed the DHA formula. They had significantly (P < 0.05) smaller gains in weight, length, and head circumference between study enrollment to 59 weeks PMA than those fed the control formula. At 51 weeks PMA, males in the DHA group had significantly smaller head circumferences (P < 0.05) and lower FFM (P < 0.05). At 59 weeks PMA, males in the DHA group weighed less (P < 0.05), had shorter recumbent lengths (P < 0.01), smaller head circumferences (P < 0.05), and lower FFM (P < 0.01) than those fed the control formula. Energy intakes from formula (kcal/d), however, were lower at 51 weeks (P < 0.05) and 59 weeks (P < 0.05) PMA in males fed the DHA formula. Adjusted for body weight (kcal/kg/d), mean energy intakes from formula at 51 and 59 weeks PMA were not significantly different between feeding groups. The differences in recumbent length, head circumference, and FFM remained statistically significant after controlling for energy and protein intakes (P < 0.01). For all males, neither FFM nor total body fat (TBF), when expressed as a percentage of total body weight, differed significantly between feeding groups. Among females, there were no significant differences between the feeding groups in measures of growth, body composition, or energy intake. The results indicated that infant formula with fish oil containing DHA and EPA in a 5:1 ratio had a significant, negative effect on growth and body composition in males during the first 6 months of life. It is not clear why the growth deficits were limited to males and not females. The eicosanoids, bioactive metabolites of omega‐3 and omega‐6 fatty acids, may mediate several important growth hormones. The present results do not support the addition of DHA alone in infant formulas. Am. J. Hum. Biol. 11:457–467, 1999. © 1999 Wiley‐Liss, Inc.