Abstract
Insect parasitoids and their insect hosts represent a wide range of parasitic trophic relations that can be used to understand the evolution of biotic diversity on earth. Testing theories of coevolution between hosts and parasites is based on factors directly involved in host susceptibility and parasitoid virulence. We used controlled encounters with potential hosts of the Aphidius ervi wasp to elucidate behavioral and other phenotypic traits of host Acyrthosiphon pisum that most contribute to success or failure of parasitism. The host aphid is at an advanced stage of specialization on different crop plants, and exhibits intra-population polymorphism for traits of parasitoid avoidance and resistance based on clonal variation of color morph and anti-parasitoid bacterial symbionts. Randomly selected aphid clones from alfalfa and clover were matched in 5 minute encounters with wasps of two parasitoid lineages deriving from hosts of each plant biotype in a replicated transplant experimental design. In addition to crop plant affiliation (alfalfa, clover), aphid clones were characterized for color morph (green, pink), Hamiltonella defensa and Regiella insecticola symbionts, and frequently used behaviors in encounters with A. ervi wasps. A total of 12 explanatory variables were examined using redundancy analysis (RDA) to predict host survival or failure to A. ervi parasitism. Aphid color was the best univariate predictor, but was poorly predictive in the RDA model. In contrast, aphid host plant and symbionts were not significant univariate predictors, but significant predictors in the multivariate model. Aphid susceptibility to wasp acceptance as reflected in host attacks and oviposition clearly differed from its suitability to parasitism and progeny development. Parasitoid progeny were three times more likely to survive on clover than alfalfa host aphids, which was compensated by behaviorally adjusting eggs invested per host. Strong variation of the predictive power of intrinsic (body color) and extrinsic traits (symbionts, host plant), indicate that host variables considered as key predictors of outcomes strongly interact and cannot be considered in isolation.