Floral evolution as a figment of the imagination of pollinators.

Schaefer and Ruxton recently reviewed the mechanisms involved in plant deception, with an emphasis on pollination [1]. They proposed that generalised food deception evolves if plants exploit the innate preferences of pollinators, and thus represents a form of exploitation of perceptual biases (EPB). They contrast this with floral mimicry, which is viewed as a distinct phenomenon that might originate with EPB before selection hones the resemblance between a mimic and a specific model. We argue that EPB in the form of pre-existing bias is not limited to deceptive plants, but rather drives the evolution of floral traits of animal-pollinated plants in general. Further, in our view, EPB is central to the evolution of floral mimicry, which differs from generalized mimicry only in the exploitation of specialised pollinators, leading to a close resemblance between mimics and specific models. Schaefer and Ruxton highlight the importance of preexisting biases in pollination through the EPB model. We agree that EPB is an important mechanism for understanding the evolution of floral traits, but suggest that it is more widely applicable to floral evolution, and that it cannot explain the evolution of deceptive flowers. The prerequisites for EPB to select for particular plant signals are: (1) pollinators having specific innate preferences; (2) plants being limited in their reproductive success by access to pollinators (i.e. increased pollinator attraction increases plant fitness); and (3) insect perceptual systems predating plant signals, thus the evolution of plant signals exploits pre-existing sensory preferences in the pollinators. There is ample evidence for all three conditions across plant and pollinator lineages. Several studies have shown that plants are often pollinator-limited in their reproductive success [2]. Phylogenetically informed approaches to the evolution of sensory systems in insects and the corresponding signals in plants indicate that insect vision and olfactory systems predated the evolution of floral colour and scent in angiosperms [3,4]. Examples of floral signals influenced by preexisting pollinator bias include the convergent evolution of red colouration in bird-pollinated flowers, which may result from better detection of the colour red in birds [5], possibly because red serves as an intra-specific communication signal in birds [6]. Floral guides (stripes, dots) might have evolved under innate preferences of bees for radiating stripes, dark centres, and peripheral dots [7]. Yellow spots on flower petals are thought tomimic pollen, selected for by