Abstract
We comment on a recent paper by Francis and Currie that reports a globally consistent relationship between plant family diversity in large geographic grid cells and factors in the physical environment. This relationship was interpreted as demonstrating local ecological determination of diversity, leaving little room for regional and historical effects. We suggest that using family richness to describe patterns of diversity reduces regional effects owing to the lower endemism of plants at the family compared with the species level. Regardless of the taxonomic level of analysis, we further point out that unless regional effects are tested explicitly, statistically significant differences in diversity between regions can be overlooked. Finally, even correlations between diversity and physical conditions of the environment can have primarily historical and evolutionary origins when lineages diversify within ecological zones of origin and spread more slowly to diverse environments. Thus, local ecological determinism should not have primacy over evolutionary-historical explanations for patterns of diversity between or within regions. Ecologists must evaluate patterns in local diversity within historical and geographic, as well as ecological, contexts.