In this article, we examine the details and dating of the mergers of Latin ı, e into e ̣ and u˘, o into o ̣. While these mergers have long since been recognised as a matter of historical fact within the broader development of Latin into Romance, their micro-level mechanics present several problems. By re-examining the available evidence from various empirical and theoretical perspectives, we argue that the mergers were the result of two phonological processes: the rise of peripherality distinctions in the vowel system on the one hand and the loss of contrastive vowel length in the fifth century AD on the other. 1 The authors wish to thank Andreas Opfermann and Michael Mazzola for reading and commenting drafts of this article. We also wish to thank the editor, James Clackson, and the two anonymous reviewers for useful suggestions and corrections. All remaining errors are our own.