Orthogonal fitness benefits of nitrogen and ants for nitrogen-limited plants in the presence of herbivores.

Predictable effects of resource availability on plant growth-defense strategies provide a unifying theme in theories of direct anti-herbivore defense, but it is less clear how resource availability modulates plant indirect defense. Ant-plant-hemipteran interactions produce mutualistic trophic cascades when hemipteran-tending ants reduce total herbivory, and these interactions are a key component of plant indirect defense in most terrestrial ecosystems. Here we conducted an experiment to test how ant-plant-hemipteran interactions depend on nitrogen (N) availability by manipulating the presence of ants and aphids under different N fertilization treatments. Ants increased plant flowering success by decreasing the densities of herbivores, and the effects of ants on folivores were positively related to the density of aphids. Unexpectedly, N fertilization produced no changes in plant N concentrations. Plants grown in higher N grew and flowered more, but aphid honeydew chemistry stayed the same, and neither the density of aphids nor the rate of ant attraction per aphid changed with N addition. The positive effects of ants and N addition on plant fitness were thus independent of one another. We conclude that N was the plant's limiting nutrient and propose that addition of the limiting nutrient is unlikely to alter the strength of mutualistic trophic cascades.

Plant-derived differences in the composition of aphid honeydew and their effects on colonies of aphid-tending ants.

In plant-ant-hemipteran interactions, ants visit plants to consume the honeydew produced by phloem-feeding hemipterans. If genetically based differences in plant phloem chemistry change the chemical composition of hemipteran honeydew, then the plant's genetic constitution could have indirect effects on ants via the hemipterans. If such effects change ant behavior, they could feed back to affect the plant itself. We compared the chemical composition of honeydews produced by Aphis nerii aphid clones on two milkweed congeners, Asclepias curassavica and Asclepias incarnata, and we measured the responses of experimental Linepithema humile ant colonies to these honeydews. The compositions of secondary metabolites, sugars, and amino acids differed significantly in the honeydews from the two plant species. Ant colonies feeding on honeydew derived from A. incarnata recruited in higher numbers to artificial diet, maintained higher queen and worker dry weight, and sustained marginally more workers than ants feeding on honeydew derived from A. curassavica. Ants feeding on honeydew from A. incarnata were also more exploratory in behavioral assays than ants feeding from A. curassavica. Despite performing better when feeding on the A. incarnata honeydew, ant workers marginally preferred honeydew from A. curassavica to honeydew from A. incarnata when given a choice. Our results demonstrate that plant congeners can exert strong indirect effects on ant colonies by means of plant-species-specific differences in aphid honeydew chemistry. Moreover, these effects changed ant behavior and thus could feed back to affect plant performance in the field.