Many modern paleobiological analyses are conducted at the generic level, a practice predicated on the validity of genera as meaningful proxies for species. Uncritical application of genera in such analyses, however, has led—perhaps inadvertently—to the unjustified reification of genera in an evolutionary context. While the utility of genera as proxies for species in evolutionary studies should be evaluated as an empirical issue, in practice it is increasingly assumed (rather than demonstrated) that genera are suitable proxies for species. This is problematic on both ontological and epistemological grounds. Genera are arbitrarily circumscribed, non-equivalent, often paraphyletic, and sometimes polyphyletic collections of species. They are useful tools for communication but have no theoretical or biological reality of their own and, whether monophyletic or not, cannot themselves operate in the evolutionary process. Attributes considered important for understanding macroevolution—e.g., geographic ranges, niche breadths, and taxon durations—are frequently variable among species within genera and will be inflated at the generic level, especially in species-rich genera. Consequently, the meaning(s) of results attained at the generic level may not “trickle down” in any obvious way that elucidates our understanding of evolution at the species level. Ideally, then, evolutionary studies that are actually about species should be pursued using species-level data rather than proxy data tabulated using genera. Where genera are used, greater critical attention should be focused on the degree to which attributes tabulated at the generic level reflect biological properties and processes at the species level.