Eight-hundred fifty-five newborns were enrolled in a prospective study between 1978 and 1982 and then followed through school age. To determine whether the mode of infant feeding affected developmental scores or school grades, prospective data were collected on how the children were fed; 788 of the children had Bayley tests at 6 months, 720 at 12 months, 676 at 18 months and 670 at 2 years. McCarthy testing was done on 645 children at 3 years, 628 at 4 years and 636 at 5 years. Testers were not specifically blind to feeding method. The families provided report cards from grade 3 or higher for 366 children. There were statistically significant but small increases in scores among breast-fed children on at least some subscales of the Bayley and McCarthy at all time points from 2 years through 5 years and slightly higher English grades on report cards in both crude analyses and in multivariate analyses that allowed adjustment for the most plausible confounding variables. We conclude that, in a volunteer, 95% white sample of middle class children, those breast-fed scored slightly better than those bottle fed; the effect is small but still detectable at school age.