Plant functional traits, especially leaf traits, are accepted proxies for ecosystem properties. Typically, they are measured at the species level, neglecting within-species variation. While there is extensive knowledge about functional trait changes (both within and across species) along abiotic gradients, little is known about biotic influences, in particular at local scales. Here, we used a large biodiversity-ecosystem functioning experiment in subtropical China to investigate intra-specific trait changes of 16 tree species as a response to species richness of the local neighbourhood. We hypothesized that because of positive complementarity effects, species shift their leaf traits towards a more acquisitive growth strategy, when species richness of the local neighbourhood is higher. The trait shift should be most pronounced, when a focal tree’s closest neighbour is from a different species, but should still be detectable as a response to species richness of the directly surrounding tree community. Consequently, we expected that trees with a con-specific closest neighbour have the strongest response to species richness of the surrounding tree community, i.e., the steepest increase of acquisitive traits. Our results indicate that species diversity promoted reduced competition and complementarity in resource use at both spatial scales considered. In addition, the closest neighbour had considerably stronger effects than the surrounding tree community. As expected, trees with a con-specific nearest neighbour showed the strongest trait shifts. However, the predicted positive effect of local hetero-specificity disappeared at the highest diversity levels of the surrounding tree community, potentially resulting from a higher probability to meet a strong competitor in a diverse environment. Our findings show that leaf traits within the same species vary not only in response to changing abiotic conditions, but also in response to local species richness. This highlights the benefit of including within-species trait variation when analysing relationships between plant functional traits and ecosystem functions.

We made use of the Tree Species Pairs (TSP) design within the BEF-China experiment (Site A) to measure SLA (specific leaf area), LDMC (leaf dry matter content), leaf C (carbon), leaf N (nitrogen), leaf C:N (carbon to nitrogen ratio), leaf Mg (magnesium), leaf Ca (calcium), leaf K (potassium) and leaf P (phosphorus). For each of the 432 studied trees we collected leaves within the interaction plane of the TSP, at 3 to 5 different heights along the crown. At each height, we cut 3 leaves for which we measured the near-infrared spectrum. Together with the trait values of additional samples processed in the laboratory, we predicted the TSPs' trait values from each single spectrum. Here, values are averaged at the tree level (1 mean trait value per tree for each trait). For detaileds methods, see Material & methods of the corresponding paper.

BEF-China main experiment : Site A Xingangshan, Jiangxi Province, Dexing County, China 29°080 29°110N, 117° 900 117°930E (geoname ID 7651294)

We used leaf traits of the 16 tree species of Site A. A full list of the species can be found in the 'Species' column of the dataset, and in Table A1 of the supplementary material accompanying the associated paper.

We analysed the effects of local neighbourhood richness (i.e, the species richness of the ten trees directly surrounding a TSP) and TSP diversity (i.e, mono- or hetero-specific TSPs) on nine structural and biochemical leaf traits.