An inverse relation between IQ and delinquency has been well established, but the direction of effect remains to be specified. Differing explanatory accounts of the relation were empirically examined in the present study using data on 13-year-old boys involved in a high-risk longitudinal study. Accounts that interpreted the relation as spurious or that posited that delinquency-related factors lead to low IQ scores received no support; findings were most consistent with the hypothesis that the direction of effect runs from low IQ to delinquency. The IQ-delinquency relation was robust after race, class, and observed test motivation were controlled statistically. Additionally, the effect of IQ was mediated by school performance for Black youth but not for White youth.