Aim: Recent studies have found that local-scale plots measured through time exhibit marked variation in the change in species richness. However, the overall effect often reveals no net change. Most studies thus far have been agnostic about the identities of the species lost/gained, as well as the processes that may lead to these changes. Generalist traits may be crucial in allowing species to colonize new plots or remain resilient in situ, whereas environmental filtering may remove specialists. We test whether plots are changing in species richness, whether they are becoming more similar (i.e. becoming homogenized) through time, as well as whether several generalist traits can predict gains or losses from local plots. Location: Canada Time period: 1945-2015 Major taxa studied: 265 species of butterflies Methods: We measured i. species richness change and ii. pairwise beta diversity across 96 well-sampled 10x10 km plots across Canada between two time periods: 1945-1975 and 1985-2015. We looked at the effects of wingspan, mobility, diet breadth and range size on the number of grid cells each species gained and lost between time periods. Results: We observed a slight increase in plot-level species richness, and that these communities are becoming homogenized through time. We note that most butterfly species in Canada have large North American ranges, but the widest-ranged species are better able to colonize new plots than narrower-ranged species, but also experience higher frequencies of local extinctions. In sum, the median range size of species within a plot increased through time. Main conclusions: We highlight that, even when local species richness exhibits very little change, other potentially important biodiversity changes, such as geographic homogenization due to the colonization dynamics of already widely distributed species, can occur. Such patterns can reconcile observed global losses of species with the simultaneous lack of change in local diversity.