Purpose: This study examines the density of polluting industry by neighborhoods in Baltimore over the long term, from 1950 to 2010, to determine if high pollution burdens correspond spatially with expected demographic and housing variables predicted in the environmental justice literature. For 1960For -1980 we use data on heavy industry from Dun and Bradstreet directories and for 1990-2010 the US EPA's Toxics Release Inventory to calculate a Hazards Density Index. Drawing on the decennial censuses for 1960-2010, we populate census tracts from corresponding years with data on race, ethnicity, educational attainment, income, and housing tenure.Findings: Density of polluting industry is positively correlated with low-income neighborhoods and renter-occupied housing in 1960 and by 2010 with white, Hispanic, and low educational attainment populations. In general, over time density of polluting facilities shifts from an association with wealth to race and ethnicity while educational attainment remains a significant variable throughout. This study confirms earlier analyses on Baltimore that white neighborhoods are more likely than African-American neighborhoods to contain polluting facilities but reveals for the first time that educational attainment is also significant. The paper concludes with a discussion of the Baltimore Sustainability Plan and its weak efforts to address persistent environmental injustices.Keywords: environmental justice, Baltimore, longitudinal, Toxics Release Inventory, Hazards Density Index, sustainability Highlights Environmental inequity is persistent over time in Baltimore Over time, density of polluting facilities shifted from associations with poverty and wealth to race and ethnicity Educational attainment strongly correlates with density of polluting facilities over time In 2010, African Americans lived in neighborhoods with fewer polluting industries than whites Sustainability plans do not adequately combat entrenched environmental inequities This is an author-produced, peer-reviewed version of this article. The final, definitive version of this document can be found online at Cities, published by Elsevier Copyright restrictions may apply.