It has been shown that lower birthweight is associated with lower IQ, but it remains unclear whether this association is causal or spurious. We examined the relationship between birthweight and IQ in two prospective longitudinal birth cohorts: a UK cohort of 1116 twin pairs (563 monozygotic [MZ] pairs), born in 1994-95, and a New Zealand cohort of 1037 singletons born in 1972-73. IQ was tested with the Wechsler Intelligence Scales for Children. Birthweight differences within MZ twin pairs predicted IQ differences within pairs, ruling out genetic and shared environmental explanations for the association. Birthweight predicted IQ similarly in the twin and nontwin cohorts after controlling for social disadvantage, attesting that the association generalized beyond twins. An increase of 1000 g in birthweight was associated with a 3 IQ point increase. Results from two cohorts add to evidence that low birthweight is a risk factor for compromised neurological health. Our finding that birthweight differences predict IQ differences within MZ twin pairs provides new evidence that the mechanism can be narrowed to an environmental effect during pregnancy, rather than any familial environmental influence shared by siblings, or genes. With the increasing numbers of low-birthweight infants, our results support the contention that birthweight could be a target for early preventive intervention to reduce the number of children with compromised IQ.