Identifying whether suicides in a region are due to characteristics of the residents living there or to some enduring feature of the region is difficult when using cross-sectional studies. To distinguish these factors, we compared the suicides of a region's residents with people who were temporarily visiting the region. Using U.S. death records from 1973-2004, we focused on states with the highest and lowest suicide rates over this period. The high suicide region consisted of Arizona, Colorado, Montana, New Mexico, Nevada, Oregon, and Wyoming; the low suicide region consisted of Connecticut, Illinois, Massachusetts, New Jersey, and New York. For each region, we considered three groups of decedents: residents who died inside the region, residents who died outside the region, and visitors to the region. Proportionate mortality ratios were calculated for all suicides and separately for firearm suicides. In the high suicide region, visitors to and residents away from the region both had elevated suicide levels, to about the same extent as residents dying inside the region. Therefore, short-term exposure to the region and being a resident of the region each predicted suicide. In the low suicide region, the suicides of residents at home were reduced, but their suicides rose dramatically once they left the area. There was no decrease in suicides among visitors to the region. Firearm use was related to the suicide levels of each region. Overall, the results suggest that both the available means to commit suicide and the contextual features of the regions contributed to their extreme suicides. We discuss how an examination of visitors can help researchers generate novel inferences about the causes of suicide.