There is evidence that factors other than local crime rates affect levels of imprisonment in different jurisdictions. There has been relatively little research, however, concerning why significant county-level differences exist in imprisonment rates around the country. The authors examined a 1998 national sample of court data from 172 counties in order to explain differences in the use of imprisonment as a sentencing option. Results of multivariate linear regression analyses revealed that all five legal variables under examination (percentage of caseload that was violent, percentage of drug trafficking cases, percentage of felony convictions relative to Index and drug trafficking arrests, level of crime, and availability of alternative sanctions) influenced prison use. Two of the five extralegal variables, Southern region and political conservatism, were also found to influence prison use as a sentencing option. Racial composition, economic disadvantage, and urbanization, did not significantly affect the level of imprisonment. The fact that an economic threat hypothesis was not supported by the data contributes to the previous mixed findings in this line of inquiry. Future research should focus on identifying key organizational and contextual determinants that keep punishment in check. Tables, appendix, notes, and references