Restoration increases bee abundance and richness but not pollination in remnant and post‐agricultural woodlands

Abstract Human land use, including agriculture, is a leading contributor to declining biodiversity worldwide and can leave long‐lasting legacies on ecosystems after cessation. Ecological restoration is an approach to mitigate these impacts. However, little is known about how animal communities and plant–animal interactions respond to the combined effects of land‐use legacies and restoration. We investigated how restoration and agricultural history affect bee (Hymenoptera: Apoidea: Anthophila) communities and pollination function. In 27 paired remnant (no history of agriculture) and post‐agricultural longleaf pine ( Pinus palustris Mill.) woodlands, we established 4–10 1‐ha plots (126 total) and experimentally restored half of them, while the other half were left as unrestored controls. Restoration was accomplished through canopy thinning which reinstates open savanna‐like conditions. We collected bees in each plot using a combination of bowl trapping and standardized netting transects. Thinning increased bee abundance by 169% and bee richness by 110%, but agricultural land use had no effect on these variables. Bee community composition was affected by restoration and was marginally affected by agricultural history. To measure pollination function, we conducted a sentinel plant experiment in which potted black mustard ( Brassica nigra L.) plants were placed out in a subset of these sites ( n  = 10) and either bagged to exclude pollinators or left open for pollinator access. Then, we measured fruit and seed set of sentinel plants to compare pollination function among the restoration and land‐use history treatments. Seed set and fruit set of sentinel plants were higher in open than bagged plants, indicating that this model system effectively measured pollination, but we found no differences in pollination based on restoration or agricultural history. These results indicate that although pollinator communities may show clear responses to restoration that are largely independent of prior land‐use impacts, this does not necessarily translate into differences in pollination function after restoration.