Some of the Supreme Court's most famous cases - from Roe v. Wade (1973) to Brown v. Board of Education (1954)-have been decided only after being held over and argued a second time. While few cases take this path, scholars have offered no systematic account for why the Court would ever take such a tack. We develop hypotheses about when reargument is most likely to occur, and test them on all formally decided cases between 1946 and 1985. More specifically, we focus on how justices' uncertainty about case outcomes affects the Court's decision to seek reargument. Our findings demonstrate that reargument is most likely to occur when multiple levels of uncertainty are present, even when we control for other factors that have been raised as explanations for this phenomenon.