Field host-specificity of the mile-a-minute weevil, Rhinoncomimus latipes Korotyaev (Coleoptera: Curculionidae)

The safe practice of biological control relies, in part, on an accurate evaluation of a potential agent's host-specificity via testing through a filter of safety. The results of laboratory tests may differ from those obtained in open field host-specificity tests, where agents are able to use their full range of host-selection behaviors. It was hypothesized that Rhinoncomimus latipes (Coleoptera: Curculionidae), the biological control agent released against mile-a-minute weed, Persicaria perfoliata (Polygonaceae), would not feed or oviposit on nontarget plants in a two-phase, open field setting. Ten weevils were placed at the base of each of 13 test plant species in a randomized complete block design with six replicates. Weevils placed at the base of mile-a-minute weed were marked with yellow fluorescent dust, and yellow weevils were subsequently found only on mile-a-minute. Weevils placed at the base of nontarget plants (marked with red fluorescent dust) rapidly colonized mile-a-minute weed. Three hours after release, the number of R. latipes found on mile-a-minute weed was significantly higher than predicted by a random distribution of weevils on all test plants. The likelihood of finding more weevils on mile-a-minute compared to nontarget plant species was 31.0% at 3 h and increased to 96.5% at 44 h after release. Whereas prerelease studies showed feeding at low levels on 9 of the 13 plant species tested here, under open field conditions R. latipes did not feed on any nontarget plant species and dispersed from these plants. In an open field setting, where the weevil was able to use its full range of host-selection behaviors, there was no observed risk of nontarget effects for any species tested. (C) 2010 Elsevier Inc. All rights reserved.