Importantly, the structure of published Gs *did not* match the structure of G for populations evolving on Gaussian landscapes. Nor did published Gs match those seen under drift. Instead, published Gs were consistent with populations that had evolved on holey landscapes (HLs). (2/10)