Do Fungal Endophytes Result in Selection for Leafminer Ovipositional Preference?

Microorganisms can alter the direction and magnitude of plant-herbivore interactions. However, how they affect the susceptibility of the host plant and their impact on the herbivore may vary depending on the scale of the interaction such as among locations, individual plants, and parts of the plant. Positive preference-performance relationships for insects in general, and for sedentary insects such as leafminers in particular, would predict that there would be negative selection pressure against females ovipositing in low-performance sites. Thus, where the presence of microorganisms causes reduced performance, selection would favor oviposition in sites where the microbes are absent. However, this assumes that ovipositing females can detect the microbes and respond by avoidance, or that females can use cues that predict the future presence of microbes and use these predictors to respond by avoidance. We used the Cameraria sp. leafminer-Quercus emoryi system and the associated fungal endophyte species that share the leaves of the host to test the hypothesis that leafminer ovipositional preference and therefore distribution are negatively associated with endophyte distribution at scales where endophytes mediate insect antagonism and where endophyte distribution is predictable. We censused fungal endophyte distribution and abundance on >20 trees over three growing seasons at the following scales: among trees, among positions within trees, among leaves, and between and within seasons. In addition, we censused leafminer distribution and abundance and used data on leafminer distributions at different spatial scales collected over the past decade from other studies in this system to determine how they are associated with endophytes. We found that leafminers and fungal endophytes were negatively associated at two spatial scales where endophytes have been shown to affect leafminer performance negatively and where endophyte distribution and abundance were predictable. Endophytes and leafminers were not associated at one spatial scale where endophytes do not affect leafminers, and distribution and abundance was not predictable. Leafminer growth and endophyte infection were positively associated over time as both increase over the season. The results demonstrate that leafminer ovipositional preference may have been selected by avoidance of fungal endophytes and associated endophyte-mediated antagonism, but that avoidance is reliant on predictable patterns of infection.