Interpreting host-test results for classical biological control candidates: Can the study of native congeners improve the process?

When an insect candidate is identified for classical biological control, the first step is to identify its fundamental host-range. An insect’s fundamental host range, however, is typically broader than its ecological host range. This discrepancy can lead to ‘false positives’ in host testing and hamper the development of potential agents. Here, we propose a novel tool for interpreting host-range tests and identifying ‘false positives’ by studying native insects closely related to the biological control candidate. We conduct a series of laboratory and field studies comparing the fundamental and ecological host ranges of Chrysochus auratus Fabricius (Coleoptera: Chrysomelidae), a beetle native to North America, and present throughout the range invaded by pale swallow-wort, Vincetoxicum rossicum (Kleopow) Barbar. (Apocynaceae). We use the results to re-evaluate the risk associated with releasing the closely-related European beetle, Chrysochus asclepiadeus (Pallas) (Coleoptera: Chrysomelidae), a biological control candidate for V. rossicum that has raised some concerns because of no-choice feeding on North American milkweed species (Asclepias spp.) during laboratory host-testing. Laboratory and greenhouse trials here show that native North American C. auratus adults can feed and complete larval development on several native Asclepias species (fundamental host range), however in the field where both closely-related plant genera are present, this species specialized only on plants in the genus Apocynum. It appears then that Asclepias species generate ‘false positives’ for Chrysochus beetles when only the fundamental host range is assessed in the laboratory, and there is a need to re-evaluate C. asclepiadeus for potential biocontrol in North America taking into account its ecological host-plant range. We advocate for the inclusion of closely-related native congeners, where appropriate species exist, to aid in interpreting host-plant testing for potential classical biological control agents.