Plant‐mediated interactions between specialist insect herbivores of <i>Melaleuca quinquenervia</i> : Implications for management of invasive populations

Abstract Classical weed biocontrol involves release of host‐specific herbivores from the weed's native range into invaded ecosystems. Often multiple herbivores are released, assuming additive interaction between them. However, herbivory exerted by one species can modify the host quality and defence of the weed species thereby altering the damage by the other herbivore. Using an experimental community module, we studied how plant‐mediated interactions between insect herbivores influence their relative effects on plant growth, using Melaleuca quinquenervia as a model system. Native to Australia, M. quinquenervia is an invasive weed in Florida, including Everglades National Park. Two congeneric herbivores Lophodiplosis trifida and L. indentata from Australia have been introduced as biocontrol agents against this weed. We investigated if interactions mediated by M. quinquenervia influenced the effects of L. trifida and L. indentata on plant growth and nutrient dynamics, and if the order of plant–herbivore interactions altered these effects. Both L. trifida and L. indentata reduced the growth and biomass of M. quinquenervia , with L. trifida exerting a stronger suppressive effect than L. indentata . Interactions between the two herbivores were either additive or independent. The addition of L. trifida enhanced the negative impact of L. indentata on plant growth, whereas the presence of L. indentata did not affect the impacts of L. trifida . Both L. trifida and L. indentata induced similar effects on nutrient allocation. Similar changes in nutrients but differential phenotypic changes suggest that the effect of L. indentata was probably compensated for by M. quinquenervia . Nutrient allocation responses were generally additive, irrespective of the order of herbivore occurrence, and the consequences of herbivory were more pronounced in seedlings than in saplings. Practical implication: Additive or independent interactions between L. trifida and L. indentata suggest that their combined effect does not diminish the effect of either herbivore as biocontrol agents. However, if only one agent was to be prioritized (e.g. due to limited resources), L. trifida can be prioritized as it exerts a stronger suppressive effect on M. quinquenervia , relative to L. indentata . If both species are used, releasing L. trifida before L. indentata may be advantageous.