干旱区生态环境知识资源中心

Ecologists have often compared ecological communities in different areas, for example on different continents. The main interest is that the communities might be more similar in the characters of their species than expected under a null model of random species assortment. We suggest that such a null model should be based only on the species observed across the samples.

Species-level convergence and community-level convergence must be distinguished. Physical filters (limitation to growth by the physical environment) can give species-level convergence, but only biotic filtering (based on species interactions) can give community-level convergence. Matching to species and species mutual matching must also be distinguished; the process is different, but the same tests work for both.

Evolutionary convergence and ecological convergence have been distinguished in the past, but there is little value in this distinction. Even if evolutionary character separation is occurring, the selection involved must be based on ecological sorting, by elimination of genotypes that are too similar. Although ecotypes can be distinguished, the results of evolution by whole species are impossible to distinguish, by present-day observation, from those of ecological sorting. Fortunately, tests for present-day evidence of evolution also work for ecological sorting, and vice-versa.

Contrary to suggestions that community convergence is untestable, valid tests do exist, They have already produced some evidence of convergence, but evidence that community-level convergence is a common and general phenomenon remains to be produced.