Seemingly small ecological changes can have large, ramifying effects that defy expectations. Such are keystone effects in ecosystems. Phloem-feeding insect herbivores can act as keystone species by altering community structure and species interactions via plant-mediated or ant-mediated mechanisms. Plant responses triggered by phloem-feeders can disrupt tri-trophic interactions induced by leaf-chewing herbivores, while ants that tend phloem-feeders can deter or prey on other arthropods. Here, we investigate how phloem-feeding herbivores change caterpillar-parasitoid interactions on Quercus alba (white oak) trees in natural forests. We factorially manipulated the presence of phloem-feeding insects as well as ant access on Q. alba branches over multiple years and sites, and measured parasitism rates of co-occurring caterpillars. While 19.3% of caterpillars were parasitized when phloem-feeders were removed, the presence of phloem-feeders completely suppressed parasitism of caterpillars (0%). This stark pattern was consistent across the diverse community of phloem-feeders and caterpillars. Our manipulation of ant access had no effect on parasitism of caterpillars, implicating a plant-mediated mechanism. We further assessed the mechanistic hypothesis that phloem-feeders suppress plant emission of caterpillar-induced volatile compounds, which could disrupt host location behavior by parasitoids of caterpillars. Phloem-feeders indeed reduced concentrations of four volatile compounds, consistent with the putative plant volatile-mediated mechanism. Given the important role of parasitoids in controlling herbivore populations, this keystone effect of phloem-feeders offers novel insight into community dynamics in forests and potentially other terrestrial ecosystems.

See Anderson et al. 2024 Current Biology "Phloem-feeding insects creat parasitoid free space for caterpillars" for full methods on data collection and experimental design. Three datasets are provided: 1) Volatiles.xlsx provides the volatile compounds, their concentration, and branch, site, and treatment information. There is a MetaData sheet with greater detail. 2) Parasitism.xlsx provides the caterpillar data organized such that each row is an individual caterpillar. These data were assembled from individual yearly data of similar format. Otherwise, no other processing was used to assemble this data. There is a MetaData sheet with greater detail on each of the columns. 3) CaterpillarRecovery.xlsx provides the sampling recovery, parasitsm, and initial stocking values of caterpillars. This data is grouped such that each row represents the summed totals for each year, site, phloem-feeder treatment, and ant-treatment. There is a MetaData sheet with greater detail on each of the columns.