The Glasgow Duplicate Diet Study was carried out to investigate the effect of lead in drinking water on the blood-lead of infants. A stratified group of 131 mothers and babies were followed from antenatal registration until the children were 3 months of age. The infants' dietary intakes of lead were assessed by composite water sampling and by the "duplicate diet" technique. Blood-leads were measured from venous samples. Statistical analysis in this paper covers sampling variability, the dependence of diet-lead on water-lead, the relationship of blood-lead in the baby's cord to that in the mother, and the contribution of water-lead to the blood-lead (at 13 weeks) of infants that were bottle-fed. The variation associated with sampling for lead in water was taken into account when estimating the relationship between water-lead and blood-lead and when discussing how standards for their respective concentrations should be linked. The conclusion is consistent with an upper guide value of 0.05 mgl-1 for the concentration of lead in drinking water.