What determines the number of species at locations along environmental gradients (e.g., of productivity, resources, or stressors)? This is one of the older research topics in ecology, but general theories have not been developed, nor general results derived from them. We argue that a key reason for this is that neither theories nor models have  been explicit about their assumptions or the scales on which they operate. 
Results/Conclusions

Here we extend the results in Scheiner & Willig (2005) by characterizing the scales (individuals, populations, and species, and their temporal and spatial scales) on which different mechanisms can act to determine the number of species. This allows us to identify classes of models that have not yet been explored, as well as to some issues with existing models. As an example, we dissect and reconstruct the influential “available energy” model. We show that the scale on which this model can operate is much narrower than the scales on which it has been applied.