This study examined whether being breastfed predicts handedness. Two large representative samples from Britain (n=13,421) and Ireland (n=8,426) were used. Multivariate probit models are estimated, reporting the marginal effects of covariates on the probability of a child being right-handed. For both datasets, children who have been breastfed for a minimum period (between 1 month and 6 weeks) are significantly less likely to be left-handed. The results are robust to a variety of controls. Hence breastfeeding is an environmental factor that predicts handedness in a non-linear way, although the mechanism is unknown.