In his 1922 paper, ‘On the Dominance Ratio,' R. A. Fisher introduced the mathematical treatment of (what would ultimately be called) random genetic drift to the emerging field of theoretical population genetics. The received historical narrative of the origins of theoretical population genetics, largely due to William Provine, focuses on Sewall Wright's emphasis on drift's role in evolution and his disagreement with Fisher, who believed drift to be evolutionarily insignificant. This paper examines Fisher's argument for the evolutionary insignificance of drift. The standard view portrays Fisher as coming to the conclusion that drift has little to no evolutionary role because of an assumption about large population size. As I will argue, Fisher's argument is subtler than that. Ultimately, the aim of this paper is to establish what are, arguably, Fisher's landmark contributions to the modeling of drift in evolution.