Data from: Differential aphid toxicity to ladybeetles is not a function of host plant or facultative bacterial symbionts
White, Jennifer A. et al. (2017), Data from: Differential aphid toxicity to ladybeetles is not a function of host plant or facultative bacterial symbionts, Dryad, Dataset, https://doi.org/10.5061/dryad.37dv8
Herbivores often defend themselves from predation by transmitting toxic plant-produced chemicals to their enemies. Polyphagous herbivores sometimes exhibit differential toxicity when found on various host plant species, which is generally assumed to reflect variation in plant chemistry. Here, however, we provide evidence that host-associated herbivore lineages can intrinsically differ in their toxic properties. Lineages of Aphis craccivora originating from black locust (Robinia pseudoacacia) are unsuitable food for the ladybeetle Harmonia axyridis, resulting in death of both larvae and adults, whereas aphid lineages originating from alfalfa (Medicago sativa) support larval development and adult reproduction. We show that locust-origin aphids remain toxic and alfalfa-origin aphids remain non-toxic when reared on any of three legume plants (fava, alfalfa or locust). Furthermore, toxicity is not a function of the facultative bacterial symbiont Arsenophonus, which is naturally present in locust-origin aphid lineages and facilitates aphid use of locust. Experimentally cured locust-origin lineages remain toxic, and an experimentally transinfected alfalfa-origin lineage remains non-toxic to H. axyridis. Instead, Arsenophonus plays an indirect role in the distribution of toxic aphid lineages by facilitating aphid use of locust. It is the parthenogenetic coinheritance of Arsenophonus and the toxic trait that observationally correlates locust-feeding with toxicity in A. craccivora, rather than host plant chemistry per se. Our results clearly demonstrate that aphid lineages intrinsically vary in their toxic properties in a way that neither plant chemistry nor bacterial symbionts can explain. A more inclusive paradigm is needed for understanding variation in herbivore defence against predators.