The effect of host abundance on the distribution and impact of biocontrol agents on purple loosestrife (Lythrum salicaria, Lythraceae)

The spread and efficacy of biological control agents are thought to increase with the regional frequency of host populations. We tested this hypothesis by quantifying the occurrence of and damage by 2 biocontrol agents, Galerucella pusilla and G. calmariensis beetles, on 52 introduced stands of purple loosestrife (Lythrum salicaria) in Ontario, Canada. An analysis of herbarium records combined with population-genetic analysis of style morph frequencies and extensive field observations made across Ontario over a 25-y period indicate that L. salicaria populations are more densely distributed in eastern than central Ontario. However, we did not detect the control agents more frequently nor was damage by biocontrol agents higher in 14 eastern stands than 17 central stands. By comparing 20 stands where beetles had been deliberately released with 32 for which there is no recorded history of beetle release, we found that beetles had spread widely beyond documented release sites. However, plants at release sites were more heavily damaged than those at non-release sites, such that previous surveys monitoring only release sites overestimated the efficacy of control. Overall, we detected herbivory by Galerucella spp. in 88% of L. salicaria stands throughout eastern and central Ontario, and only 5% of 5957 plants sampled seem to have entirely escaped damage. We recommend an iterative post-release monitoring study to determine the broader efficacy of the control program using the baseline quantitative data presented here.