The U.S. federal judiciary is the organ of national government responsible for the authoritative resolution of disputes arising under federal law. For reasons related to its historical institutional development, the system is itself a source of conflict through the issuance of non-uniform holdings by the twelve circuit courts of appeals. Due to the regional independence of these ‘sister circuits’ and the limited supervisory capacity of the U.S. Supreme Court, a substantial proportion of circuit conflicts—or circuit splits—appear to linger indefinitely. This phenomenon makes possible the cultivation of distinct bodies of national law in the twelve regions over which the appellate courts preside. Debate centers on whether regional independence and the propensity for conflict constitutes an intolerable flaw or a beneficial feature of the U.S. federal courts. Though this debate is motivated in part by competing normative commitments, it largely turns on competing assumptions about the number of conflicts, their persistence, and the consequent costs and benefits to the legal system. In this study, I examine 151 active circuit splits across four legal subject areas—search and seizure, employment discrimination, securities, and labor law—to evaluate the quantity, duration, and legal significance of intercircuit conflicts, and test whether they are resolved in a manner that suggests judicial learning. Overall, I find little evidence that conflict serves a learning function. This study sheds empirical light on the operation of the U.S. federal court system.