Data from: Functional dissimilarity, not phylogenetic relatedness, determines interspecific interactions among plants in the Tibetan alpine meadows

The hypotheses suggesting that the nature and strength of species interactions should be determined by phylogenetic relatedness have important implications for the understanding of community structure. However, to date, there is limited empirical evidence to support them. At least two basic conditions need to be met in order to expect species interactions to be determined by evolutionary relatedness: a phylogenetic signal in the traits involved in the interactions and changes in the interactions as species are more ecologically similar. Here, we report results of a removal experiment in the Chinese Tibetan plateau in which we directly assessed if the nature and/or strength of interactions among twelve alpine meadow plant species were influenced by their phylogenetic relatedness and/or their functional dissimilarity. For each plant species, we compared its biomass production when grown alone to its biomass in presence of another species and used it as a measure of species interactions. Competition between pairs of species was more frequent than facilitation, with 60% of interactions resulting in plants producing less biomass when a second species was present. We found no effect of phylogenetic relatedness on the prevalence or intensity of competition or facilitation, presumably as none of the studied traits showed phylogenetic signal. Functional dissimilarity based on maximum plant height alone was the best predictor of both the prevalence and strength of competition and facilitation, followed by functional dissimilarity using all five functional traits. Our results pinpoint the limited capacity of phylogenetic relatedness as predictor of species interactions; underlining the limitations of using phylogenetic dispersion patterns to infer mechanisms of community assembly. On the contrary, when the right functional traits are used, functional dissimilarity among species can predict both the nature and strength of their interactions; accentuating the relevance of trait-based approaches in community ecology research