Abstract
Aim Species introductions and extinctions have reorganized the earth's biota, often leaving formerly spatially distinct assemblages more similar in species composition, a process termed biotic homogenization. The study of biotic homogenization has been almost entirely focused on the change in taxonomic similarity between assemblages through time. Here, we provide a trait‐based method for calculating functional similarity through time and compare these trends in functional attributes with those trends generated from a taxonomic perspective. Location Data were produced through computer simulation and gathered from North American Breeding Bird Survey (BBS) data and published accounts of North American birds for 10 locations across the east and west coast of the United States. Methods We simulated change in assemblages with different trait types (binary and continuous), levels of trait overlap, number of traits and species richness to determine the relationship between change in taxonomic similarity (ΔTS) and change in functional similarity (ΔFS). We also assess the relationship between ΔTS and ΔFS for bird assemblages across 10 locales in the USA between 1968 and 2008. We used simple linear regression to determine the slope and correlation between ΔTS and ΔFS and used multiple regression to assess the influence of trait overlap, number of traits, species richness and the ratio of traits to species on the relationship between ΔTS and ΔFS. Results Simulations reveal that trait redundancy governs the relationship between ΔTS and ΔFS. A decrease in trait overlap increases the slope of the regression between ΔTS and ΔFS and an increase in the ratio of traits to species in the regional pool increases the correlation. The relationship between ΔTS and ΔFS for breeding birds is comparable to simulations with low trait redundancy. Main conclusions We show that often losing or gaining species from an assemblage tells us very little about the loss or gain of function, and that this scenario most often occurs when the two assemblages have high trait redundancy. It remains to be seen how prevalent this scenario is within empirical examples; however, the implications for the continued delivery of ecosystem functions in the face of species introductions and extinctions are large.