The effect of blood-lead on children's ability and attainment was investigated in a sample of 855 boys and girls aged 6-9 years from eighteen primary schools within a defined area of central Edinburgh. The geometric mean blood-lead value was 10.4 micrograms/dl. In a stratified subsample, 501 children completed individual tests of cognitive ability and educational attainment from the British Ability Scales (BAS). An extensive home interview with a parent was also done. Multiple regression analyses showed a significant negative relation between log blood-lead and BAS combined score, number skills, and word reading when thirty-three possible confounding variables were taken into account. There was a dose-response relation between blood-lead and test scores, with no evidence of a threshold. The size of the effect was small compared with that of other factors. Lead at low levels of exposure probably has a small harmful effect on the performance of children in ability and attainment tests.