The effects of flooding,plant traits,and predation on purple loosestrife leaf‐beetles

Classical biological weed control is based on the premise that introducing specialized natural enemies from the native range re‐establishes herbivore control of plant invaders, ultimately leading to negative population growth rates. Evidence from past biocontrol programs suggests that herbivores are not solely responsible for shaping plant demography. Diverse environmental conditions in the introduced range may not only affect demography, but also influence top‐down control of target plants. We investigated how flooding affects impacts of predators (top‐down) and plant quality (bottom‐up) on performance of two leaf‐beetles, Galerucella calmariensis L. and Galerucella pusilla Duftschmid (Coleoptera: Chrysomelidae: Galerucini), released in North America as biocontrol agents of purple loosestrife, Lythrum salicaria L. (Lythraceae). Predation and flooding regime have been linked to low leaf‐beetle recruitment at sites where insects failed to attain outbreak populations. Predator exclusion experiments at adjacent flooded and non‐flooded sites indicated a positive effect of flooding on leaf‐beetle survival for all developmental stages, whereas predator exposure had little effect. There was no difference in predation rates at sites with successful or failed purple loosestrife control, questioning the importance of predation in limiting growth and impact of these biocontrol agents’ populations. Effect of flooding on purple loosestrife quality was evaluated in a common garden study where plants were grown under different flooding treatments. Plants grown in flooded soil had higher water content and lower tannic acid concentration than plants grown in well‐drained soil. Consistent with field observations, leaf‐beetle oviposition rate and survival were higher on flooded plants. Results indicate that both bottom‐up and top‐down forces operate on Galerucella populations, yet their relative strength is mediated by flooding regime. Ignoring intricacies of plant‐herbivore and trophic interactions in the introduced range appears to be a major handicap for the improvement of weed biocontrol programs.