We examine the London congestion charge introduced in 2003 and demonstrate significant reductions in a number of pollutants relative to controls. We even find evidence of reductions per mile driven suggesting amelioration of a congestion externality. Yet, we find a robust countervailing increase in harmful NO2 likely reflecting the disproportionate share of diesel vehicles exempt from the congestion charge. This unintended consequence informs on-going concern about pollution from diesel based vehicles and provides a cautionary note regarding substitution effects implicit in congestion charging schemes.