I estimate localization and urbanization economies with a methodology and data different from those of Henderson and Nakamura. After demonstrating the problems associated with using a single agglomeration variable, I present my estimates. As Nakamura found for Japanese cities, I find evidence of both localization economies and urbanization economies but I also find evidence of urbanization diseconomies in some industries, as did Henderson. My evidence for localization and urbanization economies in American cities is stronger than Henderson's evidence. Finally, I show that the pattern of agglomeration economies by industry is consistent with observed patterns of industrial location.