Although the geographic range is a fundamental unit of analysis for many macroecological and biogeographical studies, as a representation of the spatial distribution of individuals it is clearly a scale-dependent abstraction. As any amateur naturalist realizes, a species is not guaranteed to be present at every point within the range delimited by a field guide. The degree to which a species is predictably encountered within its range varies tremendously across species. Using paired datasets on species ranges and survey data for birds, butterflies, and small mammals, I plan to quantify the variation in range occupancy across species, and identify the primary species level traits (e.g. body size, population density, habitat specialization) that explain this variation. The porous nature of species¿ ranges also has consequences for community level analyses. The overlaying of range maps leads to an overestimate of species richness for local communities, as well as to the perceived coexistence of species that may not actually coexist at a biologically meaningful spatial scale. I also propose to examine the degree to which local community composition as identified by surveys represents a non-random subset of the regional pool (as defined by range maps) using simulation analyses. collapse