Pollinator experience and the evolution of flowering time

Bees forage in a positively frequency-dependent manner (over-visiting the common flower type in mixed arrays, all else being equal), potentially putting rare flower types at a reproductive disadvantage. In principle, this could slow the rate of floral evolution, because new mutant flower types are initially uncommon. New flower types may suffer an additional reproductive cost if there is a lag in pollinator acceptance of a new resource. The evolution of earlier flowering time (e.g., in response to climate change) seems particularly likely to be constrained because early-flowering mutants are both rare and (necessarily) unfamiliar to pollinators. Using laboratory-reared bumble bees foraging on arrays of artificial flowers, we simulated the phenological progression of a two-species plant “community” by gradually introducing flowers of an unfamiliar colour into an array of a familiar colour. Most bees were resistant to visiting the novel flower type, sometimes returning to the colony without a full honey-crop in preference to investigating a flower of an unfamiliar colour. Rare, novel flower types typically obtained no pollinator visits under our experimental conditions, despite containing rewards equal to those in flowers of the familiar type. A simulation model of the evolution of flowering phenology in an annual plant, given abiotic selection for earlier flowering and the type of lagged frequency-dependent pollinator visitation demonstrated in the laboratory, suggests that such strong aversion to novelty can impose a constraint on flowering time evolution and lead to plant population declines or extinction. Whether populations of real plants in fact experience such a constraint is unclear.