The authors examined the effects of maternal intrauterine undernutrition on offspring birth weights in a cohort of women born between August 1944 and April 1946 in Amsterdam, The Netherlands. This period included the Dutch Hunger Winter, a war-induced famine. Undernutrition was defined separately for each trimester of pregnancy as an average supply of less than 1,000 calories per day from government food rations. For maximum control of potential maternal confounding factors related to offspring birth weight, the authors performed a within-family analysis, including 437 families with two siblings and 107 families with three siblings born between 1960 and 1985. As in other studies of the famine, maternal birth weight itself was decreased after third trimester intrauterine exposure but not after first trimester exposure. The expected increase in offspring birth weights with increasing birth order was not seen after maternal intrauterine exposure in the first trimester of pregnancy. In this group, second born infants weighed, on average, 252 g less at birth than their firstborn siblings (95% confidence interval (CI) -419 to -85), and thirdborn infants weighed 419 g less (95% CI -926 to 87), even after adjustment for trimester of maternal intrauterine exposure, maternal birth weight, smoking during pregnancy, and sex of infants in the sibling pairs. Additional adjustment for the birth weight of the elder sibling did not materially change this abnormal pattern. There were no abnormal patterns in offspring birth weights after maternal intrauterine exposure in the second or third trimester of pregnancy. The study outcomes could not be explained by other selected determinants of birth weight, by lack of control for socioeconomic status, or by loss to follow-up of the 1944-1946 birth cohort. This study suggests that there may be long-term biologic effects, even into the next generation, of maternal intrauterine undernutrition which do not correspond to the effects on the mothers' own birth weights.