Abstract
We tested the effect of the native container-dwelling predator Toxorhynchites rutilus on the codominant container-dwelling mosquitoes: native Aedes triseriatus and invasive Aedes japonicus. We established two predator treatments (predator, no predator) by removing T. rutilus from all containers, and stocking T. rutilus larvae (1/3.5 liters) in the predator treatment. Removal and stocking was repeated every 3 d and established significantly different predator abundances in both large and small containers. Repeated-measures analysis of variance (ANOVA) on standard samples showed larvae+pupae/liter of A. japonicus was greater without versus with predation, and this difference increased across samples. In contrast, repeated-measures ANOVA showed larvae+pupae/liter of A. triseriatus was statistically indistinguishable for predation treatments and was greater in small versus large containers. Thus, predation reduced invasive A. japonicus while having no detectable effect on A. triseriatus larvae and pupae. A final destructive census of pupae showed that predation reduced pupae/liter of both species, but this effect was greater and more consistent across container sizes for A. japonicus. Predator effects on abundances were not products of the nonlethal effect of predator avoidance by ovipositing females, as T. rutilus presence did not lead to reduced egg inputs by either Aedes, nor by Aedes spp. as a group. Effects of predation thus are best explained by differential success of developing larvae due to the greater lethal effect of T. rutilus on A. japonicus than on A. triseriatus. Thus, this system is consistent with the hypothesis that native predators can limit success and potential impacts of invasive mosquitoes.