Background/Question/Methods The potential of macroecology to reveal the broad responses of species and communities to global change is widely recognised. At smaller, regional scales, however, important changes in ecosystems may not be detected by traditional macroecological analyses focused on relationships between distribution, abundance and body size between species. In such analyses, species are generally treated as homogeneous units, with the implicit assumption that all individuals conform to the species-level mean. While the extent of human impacts on many marine systems is such that they leave a clear footprint on such broad-brush interspecific macroecological relationships, in less disturbed systems variation within species may become more significant. For instance, it is at the individual level that ecological interactions (e.g. trophic interactions) occur, and especially in aquatic systems the role an individual plays in such interactions can be determined more by its size than by its specific identity. Equally, the degree to which a fish is targeted by a fishery will depend on its size as well as on its species. Here, we develop a method for incorporating intraspecific variation in body size into a sensitive macroecological index of human impacts on lightly exploited reef fish communities. We use data on the size of all individual fish of 184 species recorded in visual surveys of coral reefs surrounding 13 islands in the Lau group, Fiji, to compare species-level and size-structured macroecological relationships. These reefs are subject to subsistence fishing only, but the degree of fishing varies spatially as a function of human population density. Results/Conclusions Even in these lightly exploited fish communities, we find clear size-structured effects of fishing, evidenced by a truncation of the community size spectrum as a function of fishing intensity both within and across species. The strong size structuring of these communities is revealed by the relationship between site occupancy and local population density, which is much more pronounced when size classes rather than species identity are used as the unit of analysis. The species-level relationship between occupancy and density has been shown in other systems to be a sensitive indicator of both exploitation and extensive habitat modification. We find no such species-level effect, but subsistence fishing has a strong effect on the size-structured relationship. Low intensity human activity may not therefore leave a detectable signal on simple species-level analyses, but incorporating individual variation into a macroecological approach can reveal profound effects on ecosystem structure and function.