--- - |2 Objective In this article, we explore the deaths in office of U.S. senators between 1919 and 2015, examining both historical trends at the aggregate level and at the individual level searching for partisan and other patterns in mortality rates. Methods We employ Cox proportional hazard models to examine the effects of factors such as age, tenure in office, electoral factors, and legislative engagement, as well as partisanship and ideology. Results Notably, we find no significant partisan or ideological effects. When we examine the parties separately, we find that other institutional factors (tenure in office, vote share, bill sponsorship) matter for Democrats, while expectation of who would replace them in office matters for Republicans. Conclusion Our findings contribute to previous work on legislative turnover in Congress by taking the first step to analyze deaths among U.S. senators from almost the entire history of the elected Senate. - Social Science Quarterly, EarlyView.