Main Experiment: Ant-wound-interactions Main Experiment

Usage Rights

This data is Free for members.

Please contact Michael Staab ( before analyzing the data or using them as covariates in synthesis. There might be updates on the taxonomy of the ants and on the general framework of the data.

Dataset Abstract

Ants are abundant and functionally important arthropods in tropical and subtropical forests. The vast majority of vegetation-foraging ants were thought to be predacious or omnivorous, until studies of stable isotopes showed that those ants are essentially feeding as herbivores, mostly by visiting extrafloral nectaries and by tending Hemiptera for honeydew. Additionally, there are anecdotal records of wound feeding, i.e. the consumption of plant sap from wounds created by sucking or chewing herbivorous insects. However, to our surprise, no study has investigated and quantified wound feeding at the level of entire ant communities.
We conducted over several years an exhaustive survey for wound-feeding ants in the experimental tree plantations of the BEF-China experiment, located in subtropical South-East China. In total, we observed wound-feeding ants on 23 out of 40 planted tree species. There was a strong bias towards Fagaceae species; 90% of all observations occurred on eleven species of Fagaceae, albeit Fagaceae species accounted for less than 50% of all surveyed tree individuals. Almost all trophobiotic ant species known from the study area were also found at wounds. Our data indicate that wound feeding is an unspecific, opportunistic, and facultative behavior. Albeit less common by an order of magnitude when compared to trophobioses, feeding on wound sap might be an important but underestimated food source that contributes to fuel the diversity and abundance of vegetation foraging ants, for example in cases when herbivorous insects are abundant or trophobiotic Hemiptera are scarce. Interestingly, increasing tree diversity decreased the specialization while increasing the complementarity of plant-ant networks. Both effects were independent of network size and sampling effort, and support recent theoretical and experimental studies, which demonstrated a structuring and stabilizing effect of tree diversity on species interaction networks.

Dataset Design

Data were collected by visually scanning 20 leaves per tree individual for all trees in the extended core areas of all plots (plots with conifers, free succession, and plots with shrubs were not sampled). Site A was assessed once in 2011 and twice in 2014, site B once in 2013 and twice in 2014.

Spatial Extent

BEF China Main Experiment, near Xingangshan, Jiangxi

Temporal Extent

Mai 2011 - October 2014

Taxonomic Extent

Ants foraging for wound sap

Data Analysis

The data unit of all data is the occurrence of an ant-wound interaction, i.e. every interaction between an ant and a plant wound observed on an individual tree was only counted once, regardless of the number of ant individuals involved. Hence data are 'occurrences', i.e. whenever an ant interacted with a wound it was counted as a '1'. Raw abundance data (which are not very meaningful) are available from Michael Staab (

Data columns available in the raw data part of this dataset

Name of each plot, coded with letters and numbers in Site A; Datagroup description: Reasearch plots of the Biodiversity - Ecosystem functioning experiment (BEF-China). There are three main sites for research plots in the BEF Experiment: Comparative Study Plots (CSP) in the Gutianshan Nature Reserve, having a size of 30x30m^2, measured on the ground. Main Experiment plots have a size of 1 mu, which is about 25x25m^2 in horizontal projection. Pilot Study Plots have a size of 1x1 m^2. Research plots on the main experiment have a "p" in front of their IDs and then a 6 digit code: Plots in the main sites A and B are named according to their position in the original spreadsheet, in which they were designed. They consist of 6 digits: _1st digit_: Site (1:A, 2:B), _digit 2and3_: southwards row: as in spreadsheets the rows are named from the top to the bottom; _digit 4 and 5_: westward column: as in the original spreadsheet, but the letters are converted to numbers (A=01, B=02); _6th digit_: indicator, if the plot has been shifted a quarter mu. Example: "p205260": "p" means that this is a plot that is specified. "2" means, that we are at site B. Now the coordinates of the south - west corner: "0526". Since "e" is the fifth letter of the alphabet, this is Plot E26. The last digit "0" means that this plot was not moved by a quarter of a Mu, as some sites in Site A. The 6th digit can also indicate the subplot within the plot. "5", "6", "7", "8" indicate the northwest, northeast, southeast, and southwest quarter plot respectively.
Data group: Plot name Main experiment (letter, number code)
Keywords: location
Name of the study site of the main experiment
Data group: Experimental Site
Keywords: Main Experiment, location
Planned tree species richness according to design.; Datagroup description: Planted number of tree species per plot
Data group: Taxonomic biodiversity
Keywords: biodiversity
Position of a given tree individual per plot given in XXYY; Datagroup description: There are 20x20 trees per plot. Tree_label marks the position of each tree in a coordinate system starting at 0000.
Data group: Position within Main experiment plots
Keywords: Main Experiment, location
Taxonomic name of tree species.
Data group: Scientific plant species name
Keywords: species, taxon
Betula luminifera
Castanopsis eyrei
Castanea henryi
Alniphyllum fortunei
Castanopsis fargesii
Taxonomic name of ant species
Data group: Ant species name
Keywords: species, taxon, arthropods, ant
Camponotus albosparsus
Dolichoderus sibiricus
Camponotus vitiosus
Crematogaster cf. nawai
Crematogaster cf. rogenhoferi
morphospecies of the respecitve ant
Data group: Nominal taxon
Keywords: arthropods, ant, trophic interactions
Prenolepis sp.6
Prenolepis sp.5
Nylanderia sp.3
Number of ant-wound interactions observed.; Datagroup description: Number of ant-wound interactions observed.
Unit: occurrences
Data group: Trophobiosis (count)
Date the interaction was observed
Unit: days
Data group: Date time information
Sampling round
Data group: Helper