Ranking the host range of biological control agents with quantitative metrics of taxonomic specificity

Biological control of pest insect, plant and microbial organisms with their natural enemies (biological control agents) is an important part of sustainable crop protection in Canada. Because of the need to prevent any potential negative impacts of importation biological control on native biodiversity, evaluating the specificity (host range) of natural enemies has become a central issue for the application of biological control programs against pest organisms. Thus, when there are a number of candidate species being considered for introduction as biocontrol agents against a particular target pest, it is important to compare their relative host ranges. However there are currently no quantitative tools that are currently in widespread use to compare the relative host ranges of different candidate organisms. In this article, we propose using two quantitative metrics (taxonomic host range - STD; phylogenetic species variability - PSV) that have been previously used in other areas basic ecological research as tools to rank the relative host specificity of natural enemies used in biological pest control. These two metrics capture information not only regarding how many species a natural enemy attacks, but also how closely related they are. We first show that although these two metrics were developed independently, they are directly related mathematically and convey exactly the same information except on different numerical scales. We then apply the metrics to case studies regarding the host range of candidate biological control agents for two insects that have the potential become invasive pests of Canadian agriculture. We also provide open source code to easily allow the calculation of these metrics for any natural enemy for which taxonomic host range information is available. Finally, we discuss the range of potential uses of these metrics in applied biological control as well as important caveats and limitations that require further theoretical work to address. We conclude that while these metrics should not be considered as absolute or infallible measurements of the specificity of biological control agents, their application should encourage biological control practitioners to explicitly consider the phylogenetic component of host range when ranking prospective candidate biological control agents.