I investigated aspects of a ‘payback’ hypothesis that postulates that virus infection triggers changes in host plants that render them more attractive to pollinators. This builds on previous work, which showed that in tomato Cucumber mosaic virus (CMV) infection enhanced emission of volatile organic compounds (VOCs) that attracted bumblebees and enhanced pollination of infected plants. To test if this hypothesis is tenable with other viruses and with plants other than tomato, and to determine if pollinators might derive some advantage by visiting flowers of infected plants, I used two common bean (Phaseolus vulgaris) varieties, and three bean-infecting viruses (a CMV bean isolate, Bean common mosaic virus, and Bean common mosaic necrosis virus). Commercially produced bumblebees (Bombus terrestris) and wild bees were studied, respectively, under glasshouse conditions and in the field. My data indicates that viruses can pay back susceptible hosts by attracting pollinators through changes in host-emitted VOCs and rewarding pollinators through greater nectar quantity and sucrose concentration. The enhanced pollinator attraction correlated with a recovery in seed production in virus-infected bean plants.

Virus infection delayed the onset of flowering and decreased flower numbers, but it also caused bee-perceptible changes to flower petal colour, increased nectar volumes and nectar sucrose concentration. Changes in nectar volume/concentration are likely to encourage bee visitation. Gas chromatography-mass spectrometry revealed that headspace VOCs emitted by virus-infected plants were qualitatively distinct from those of mock-inoculated plants and that virus-infected plants emitted greater quantities of VOCs. In free-choice olfactometry assays, bumblebees displayed an innate preference for VOCs emitted by non-flowering BCMNV-infected plants and both non-flowering and flowering BCMV and CMV-infected plants over those from mock-inoculated plants. Where bumblebees showed no innate preference, as was the case for flowering BCMNV-infected plants, differential conditional assays showed that bumblebees were nevertheless able to perceive differences between the VOCs emitted by BCMNV-infected plants and mock-inoculated plants.

I examined pollination and seed production in virus-infected bean under glasshouse conditions using B. terrestris and in the Cambridge University Botanic Garden using CMV-infected plants exposed to naturally occurring bees. Under Garden conditions, I found that common carder bees (B. pascuorum) were the main bean flower pollinators, while B. terrestris and honey bees acted as nectar thieves. Under both conditions, virus-infected plants showed a recovery of seed numbers to levels similar to those from uninfected plants if pollinators were allowed access. These observations of virus-induced effects on plant-pollinator interactions support the idea that viruses may act mutualistically with plants by making infected plants more attractive to pollinators and suggest that pollinators may also derive benefits from visiting infected plants.