For more than a century, ecologists have known that the area of distribution of a species and its ecological attributes, or ecological “niche” are related. However, disentangling the details of this relationship has proven difficult.

Results/Conclusions

In this presentation, I propose that a possible avenue to clarify this relationship is by accepting that niches of species should be subdivided hierarchically into several different concepts, each better defined at different scales and defined in terms of different classes of variables in multidimensional environmental space, and that niches can best be understood by taking into account, explicitly, movements in geographic space. At coarse spatial scales (large extents and low resolutions), the Grinnellian niche can be usefully defined as subsets of an environmental space composed by values of mostly non-interactive climatic and topographic variables. At such coarse scales, details about resource utilization, habitat availability, and biotic interactions, probably “average out” (see caveat below) and the non-interactive variables likely determine the broad aspects of distributions of species. At fine scales, Eltonian niches can be defined in terms of interactive variables, where requirements and impacts both matter. At these fine scales, Eltonian processes determine the high-resolution details of the distribution. An important caveat is that the high-resolution Eltonian processes may also affect broad scale patterns. This scheme—and particularly the Grinnellian concepts—offers the advantage of interfacing well with the idea of “ecological niche modeling,” for which there are literally thousands of worked examples of applications and procedures.