Machine learning techniques can be used to identify within a large collection of digitized volumes a subset similar to a small and curated corpus of selected texts, in this case French translations of James Fenimore Cooper. As a genre the Cooperesque novel can be defined empirically, and volumes classified as Cooperesque can be sorted according to computationally determined topics. With 463,054 digitized volumes in French published between 1789 and 1914 in the HathiTrust Digital Library, we can study the nineteenth-century popular novel at scale.