We discuss claims that both uniform and mid-domain modeling approaches are equally valid in studies of species richness gradients. We argue that although they are both valid null models, they are not equally realistic in naturally bounded or virtually unbounded domains, and each model should be considered for different use in macroecology.