Incidence of EBL (blood lead > or =10 microg/dL) for children aged < or = 1.3 years in Washington, DC increased more than 4 times comparing 2001-2003 when lead in water was high versus 2000 when lead in water was low. The incidence of EBL was highly correlated (R2 = 0.81) to 90th percentile lead in water lead levels (WLLs) from 2000 to 2007 for children aged < or = 1.3 years. The risk of exposure to high water lead levels varied markedly in different neighborhoods of the city. For children aged < or =30 months there were not strong correlations between WLLs and EBL, when analyzed for the city as a whole. However, the incidence of EBL increased 2.4 times in high-risk neighborhoods, increased 1.12 times in moderate-risk neighborhoods, and decreased in low-risk neighborhoods comparing 2003 to 2000. The incidence of EBL for children aged < or =30 months also deviated from national trends in a manner that was highly correlated with 90th percentile lead in water levels from 2000 to 2007 (R2 = 0.83) in the high-risk neighborhoods. These effects are consistent with predictions based on biokinetic models and prior research.