Relative performance on test and target plants in laboratory tests predicts the risk of non-target attack in the field for arthropod weed biocontrol agents

We tested the hypothesis that quantifying the relative performance of candidate weed biocontrol agents on test and target plants during laboratory host-range testing can predict the probability of test plants being attacked in the field. For arthropod weed biocontrol agents established in New Zealand between 1929 and 2010, the probability of host use was positively correlated with relative performance for both no-choice starvation tests and no-choice oviposition tests. Moreover, multiplying together the relative performance scores for no-choice starvation and oviposition tests to create a combined risk score for no-choice tests resulted in a clear-cut threshold indicating that for scores between 0.21 and 0.33 the probability of host use (including minor spillover attack) occurring rose from close to zero to a virtual certainty. Choice oviposition test data showed a similar pattern to no-choice data, except there were two cases where host use occurred in the field on plant species that had very low relative performance scores in the choice tests. Both of these cases were associated with asynchrony between seed-feeding biocontrol agent activity and the reproductive phenology of the target plant, indicating that choice tests may be inappropriate for certain guilds of biocontrol agent that attack ephemeral plant structures. We conclude that quantitative laboratory testing data can help predict risk of non-target attack. The ability to refer to threshold relative performance scores when deciding whether an agent is safe to release has the potential to ensure fewer environmentally safe candidate biocontrol agents are erroneously rejected, thereby enhancing efficiency in the selection and approval of new agents. (C) 2014 Elsevier Inc. All rights reserved.