Abstract. 1. Chinese privet (Ligustrum sinense Lour.) was removed from riparian forests in the Piedmont of Georgia in November 2005 by mulching with a track-mounted mulching machine or by chainsaw felling. The remaining privet in the herbaceous layer was killed with herbicide in December 2006.
2. Bee (Hymentoptera: Apoidea) abundance, diversity and community similarity in the forests were measured for 2 years after shrub removal and compared with heavily invaded controls and with non-invaded forests (desired future condition) using pan traps.
3. In 2006, control plots averaged 8.8 species and 34.8 bees per plot. Privet mulching resulted in 32.5 bee species and 418.3 bees per plot, and privet felling plots had 29 species and 259 bees per plot. In 2007, control plots averaged only10 species per plot and 32.8 bees per plot, while mulched and felled plots had 48 and 38 species per plot and 658.2 and 382.5 bees per plot, respectively.
4. The bee community on untreated control plots was dissimilar from the communities on privet felling, mulched and desired future condition plots during both years; however, by 2007, desired future condition, felling and mulched plots had similar bee communities.
5. Removal of an invasive shrub provided immediate benefits for native pollinators and resulted in bee communities similar to non-invaded forests even without further restoration of native plant communities.