Theories of race-making have recognized the confusion and contradiction in state-led racial projects but have not sufficiently elaborated their unintended consequences. Focusing on the relationship between the state, racial science, and immigration policy in the early twentieth century United States, this article illustrates how practical challenges in racial projects can jeopardize and thereby eventually trigger innovations in modes of racial governance. The Dillingham Commission (1907–1911) was a Congressional investigative commission that attempted to collect comprehensive data on immigrants in order to (...) provide a scientific foundation of immigration policy. Specifically, the most powerful members of the Commission’s executive committee wanted to single out Southern and Eastern Europeans (SEEs) and portray them as “undesirable races.” Developing such a classification scheme, however, proved to be difficult, and the facts collected from the field did not support their goal, instead demonstrating that in many cases SEEs were not very different from other immigrants and their native-born counterparts. The idea for national quotas, as well as the theoretical foundation for the inclusion of SEEs, emerged in this process, not as an outcome of ideological design but as an unintended byproduct of knowledge production. I highlight the pursuit of facts-based governance and relative autonomy of middle-level managers as key factors that enabled such process. (shrink)