Lead burdens in children were studied in relation to exposure to lead from soil and from automobile traffic in Charleston, SC. Preschool aged black children exposed to a variety of soil leads and traffic volumes were recruited in a house-to-house survey. Data regarding soil, traffic, and paint exposure and demographic information also were collected. Soil lead concentrations ranged from 9 to 7,890 ppm with a daily median of 585 ppm. Children's homes faced streets with a daily volume of < 1,000 to 16,000 cars with 51% being in the high traffic recruitment area. Lead paint concentrations varied both within and between homes (0-30 mg/sq cm in porch railings, 0-51 mg/sq cm in exterior sidings). Children's blood lead levels ranged from 18 to 77 micrograms/dl with a mean of 38 micrograms/dl. Since three representations of traffic volume were available, a separate regression analysis was done for each. Soil lead was always statistically significant while traffic volume and lead paint (exterior siding and door frame) were significant when traffic was represented as high-low. Little influence of demographic or individual characteristics on blood leads was found. These results suggest that measurement of all environmental lead sources must be considered in future lead body burden studies.