The association between blood lead level and dental caries was evaluated in cross-sectional analyses of baseline data for 543 children 6-10 years old screened for enrollment in the Children's Amalgam Trial, a study designed to assess potential health effects of mercury in silver fillings. Approximately half of the children were recruited from an urban setting (Boston/Cambridge, MA, USA) and approximately half from a rural setting (Farmington, ME, USA). Mean blood lead level was significantly greater among the urban subgroup, as was the mean number of carious tooth surfaces. Blood lead level was positively associated with number of caries among urban children, even with adjustment for demographic and maternal factors and child dental practices. This association was stronger in primary than in permanent dentition and stronger for occlusal, lingual, and buccal tooth surfaces than for mesial or distal surfaces. In general, blood lead was not associated with caries in the rural subgroup. The difference between the strength of the associations in the urban and rural settings might reflect the presence of residual confounding in the former setting, the presence of greater variability in the latter setting in terms of important caries risk factors (e.g., fluoride exposure), or greater exposure misclassification in the rural setting. These findings add to the evidence supporting a weak association between children's lead exposure and caries prevalence. A biologic mechanism for lead cariogenicity has not been identified, however. Our data are also consistent with residual confounding by factors associated with both elevated lead exposure and dental caries.