A great deal of variability exists in American local government structures. It would be useful to analyze the differences in outcomes associated with different structures, but two factors stand in the way. First, there is no agreement about how to classify city governments. Second, many of the components used in existing classifications are inaccurately reported or missing in survey data. In this research note, a new classification system based on three easily determined components is developed and used to measure the current distribution of forms and variations in an expanded data set that includes almost all cities over 10,000 in population.