Scholars often extoll the benefits of knowledge accumulation. Natural experiments, however, are often thought of as idiosyncratic and one-off studies that may not therefore contribute to cumulative learning. This chapter explores this case against natural experiments. It emphasizes two key dimensions of knowledge about causal effects—generalizability and mechanisms—and underscores three empirical strategies for boosting accumulation: comparing studies in which (1) context varies but treatments and outcomes are similar; (2) different treatments are employed in the same context and with the same outcome measures; and (3) similar treatments are carried out in the same context with distinct, but related, outcomes. Surveying examples of natural experiments across different substantive areas, the authors find that scholars can leverage these strategies to foster cumulative learning. However, several features of these designs and of their use do pose barriers to understanding of generalizability and mechanism. The chapter outlines several ways in which knowledge accumulation using natural experiments can be further enhanced.