1. Grassland restoration is an important tool for conserving bee biodiversity within agricultural landscapes. Restorations foster increases in local bee abundance and α-diversity, however, these measures are insufficient for understanding if remnant communities are being conserved. We compared native bee α-diversity, β-diversity, and community composition between restored and remnant prairies in Minnesota, USA. We then investigated two potential drivers of bee community dissimilarity between restored and remnant prairies: proportion of agricultural land surrounding a restoration and differences in floral community between restored and remnant prairies.

2. We selected ten restored prairies that lie along a gradient of increasing agricultural land cover, ranging from 20–85% of surrounding land in agricultural production. We paired each restoration with a nearby prairie remnant and sampled bee and floral communities concurrently in each restoration-remnant pair. We quantified bee and forb α-diversity, community composition, β-diversity, and levels of dissimilarity between restoration-remnant pairs along the gradient of agricultural development. Additionally, we quantified differences in the community weighted mean between restored and remnant prairies for two bee traits, oligolecty and tongue length, to investigate how differences in floral community between restored and remnant prairies may influence bee community composition.

3. While bee α-diversity between restored and remnant prairies was similar, bee composition between restorations and remnants were significantly different with restorations being more homogenous than remnants. Differences in bee community composition and β-diversity were not statisticlly related to agricultural landscapes or floral community dissimilarity, however, we found a significantly higher proportion of oligolectic bees in remnant prairies. This difference in oligolectic bees was likely related to the absence of important host plants in restorations.

4. ‘Synthesis and application’: Prairie restorations should seek to provide diverse floral resources that are of similar composition to remnant prairies. Specifically, providing important floral host plants for pollen specialist bees could improve restorations’ ability to conserve prairie remnant bee communities.

This data set was collected over two years using bee sampling techniques that were standardized accross locations and detailed in the JAE manuscript titled "Differences in bee community composition between restored and remnant prairies are more strongly linked to forb community differences than landscape differences".

The five csv's in this data set are used by the R script file to recreate the analysis and plots for the JAE manuscript "Differences in bee community composition between restored and remnant prairies are more strongly linked to forb community differences than landscape differences" in the statistical program R. Files provided represent "cleaned" data ready for analysis to simply R code.