Direct and plant trait‐mediated effects of the local environmental context on butterfly oviposition patterns

Abstract

Variation in the intensity of plant–animal interactions over different spatial scales is widespread and might strongly influence fitness and trait selection in plants. Differences in traits among plant individuals have been shown to influence variation in interaction intensities within populations, while differences in environmental factors and community composition are shown to be important for variation over larger scales. However, little is still known about the relative importance of the local environmental context vs. plant traits for the outcome of interactions within plant populations. We investigated how oviposition by the seed‐predator butterfly Phengaris alcon on its host plant Gentiana pneumonanthe was related to host plant traits and to local environmental variation, as well as how oviposition patterns translated into effects on host plant fruit set. We considered the local environmental context in terms of height of the surrounding vegetation and abundance of the butterfly's second host, Myrmica ants. The probability of oviposition was higher in plants that were surrounded by lower vegetation, and both the probability of oviposition and the number of eggs increased in early‐flowering and tall plants with many flowers in the three study populations. Flowering phenology, shoot height and flower production were, in turn, related to higher surrounding vegetation. Myrmica abundance was correlated with vegetation height, but had no effect on oviposition patterns. Oviposition and subsequent seed predation by the caterpillars strongly reduced host plant fruit set. Our results show that plant–animal interactions are context‐dependent not only because the context influences the abundance or the behavior of the animal interactor, but also because it influences the expression of plant traits that affect the outcome of the interaction. The results also demonstrate that heterogeneity in environmental conditions at a very local scale can be important for the outcomes of interactions.