This data set is described in full detail in Huang et al. (2022) Journal of Ecology XXX.

In brief, we used a large tree biodiversity experiment (BEF-China), established in 2009-2010, to test whether the application of fungicide or insecticide changes observed tree species richness effects on tree growth. We used a subset of plots in which tree species numbers ranged from 1 to 8. To these plots, a factorial split-plot treatment was added in April 2014. The new treatments (I: insecticide; F: fungicide; C: untreated control) were applied to subplots located along one side of the main plots. Each subplot contained 4 × 4 = 16
trees. We further used the central 4 × 4 trees of the main plot for additional measurements (central control subplot).

Insecticide and fungicide solutions (4 L per subplot) were sprayed over tree crowns every 4 weeks, but only on days with no or very little wind. During the rainy season, application
intervals were halved to 2 weeks to compensate for more rapid leaching. The insecticide solution contained 10 mL dimethoate (an organophosphate) and 10 mL deltamethrin (a
pyrethroid). The fungicide solution contained 8 g of mancozeb (a dithiocarbamate) and 25 mL of myclobutanil (a triazole). Control subplots were sprayed with 4 L of water.

In September 2014 and 2016, we measured basal diameter and height of all trees in the subplots. We further measured the central 16 trees in each main plot because these are the trees that are censored as part of the regular inventories of all 566 BEF-China plots. Tree growth was estimated as increase in basal diameter from 2014 to 2016. We also used the product of basal diameter and height as a proxy for tree volume, and calculated volume growth as increment from 2014 to 2016. To obtain species-level basal area and volume, we added the values for all tree individuals of the same species per subplot. Community-level basal diameter and volume were than obtained by summing all the species-level values per subplot.

We quantified leaf damage in all monocultures in September 2016. We randomly chose five individuals per species from the central 6×6 trees of each main plot. Then, for each tree, we randomly picked three branches from different canopy layers and sampled seven young, fully expanded leaves per branch. The damage of these leaves was classified based on the fraction of leaf area affected (0, <5, <25, <50, <75, and >75%), separately for herbivory and fungal damage. Species mean leaf damage was calculated for each subplot by averaging the center of these damage classes over all leaves (e.g. 15% for the 5–25% class).

Data are in stored as comma-separated plain text files with utf-8 encoding. The data can be read with most software, including R (www.r-project.org). For further details of use, see the comments included  in the R scripts.