Flowers Use Physics As Well As Color And Shape To Lure Pollinators

Light focusing

When Noda et al. (1994) identified the mixta mutant of Antirrhinum majus they expected the causal mutation to affect the production of anthocyanin pigments as the flowers appeared paler than those of the wild type (Fig. 1a). However, the gene affected instead controlled the shape of the epidermal cells: in the absence of a functional MIXTA protein, the pigments remain unchanged but the cells in the petal epidermis are flat rather than the characteristic conical shape found in the wild type (Noda et al., 1994). The conical geometry (Fig. 1b) has a double effect: it acts as a lens by focusing light into the vacuole (containing the anthocyanin pigment), which increases the intensity of the colour, but it also scatters the light reflected from the mesophyll more efficiently than a flat surface. As a result, the colour saturation is increased and the petal acquires a sparkling appearance (Gorton & Vogelmann, 1996).

Is this optical trick part of the flower’s arsenal to attract pollinators? Conical cells are present on the petal epidermal surface of the vast majority of angiosperms, and generally on the section that is directly exposed to pollinators (Kay et al., 1981; Papiorek et al., 2014). Wild-type A. majus with conical cells receive more pollinator visits and produce more fruits than mutants with flat cells (Glover & Martin, 1998; Comba et al., 2000; Dyer et al., 2007) and bumblebees can perceive the difference in hue generated by cell shape (Dyer et al., 2007). Thus, by ‘playing with the light’ to modify the appearance of the flower, conical cells can generate a cue that can be detected at a distance by pollinators. This cue could help pollinators to distinguish flowers with different handling properties (see section III for role of conical cells in providing grip) (Dyer et al., 2007; Whitney et al., 2011).

Gloss

Having a smooth, lustrous surface is an efficient way of being conspicuous and gloss is another optical effect, produced by physical means, which has been described in flowers of multiple species (Fig. 1c) (Parkin, 1928; Gaisterer et al., 1999; Vignolini et al., 2012a,c; Whitney et al., 2012; Papiorek et al., 2014). Petal gloss can be defined as the specular reflection of light from the surface of the petal and this effect is maximized when the petal surface is flat (Gaisterer et al., 1999; Whitney et al., 2012). However, most flowers do not have flat cells in the visually active part of their petal epidermis and are thus not expected to be very glossy (Kay et al., 1981; Papiorek et al., 2014). Instead, gloss appears only at the tips of the conical cells commonly found on petal epidermis, generating a pattern of regularly arranged angle-dependent highlights (Papiorek et al., 2014). Alternatively, gloss can occur in defined regions of the petal, for example in the green pseudonectaries of some Solanum species.

Whether gloss acts as an attractant for insects or birds remains an open question: Gaisterer et al. (1999) found that, although gloss reduced the colour contrast of the petal of Ranunculus lingua against the background, the bright flashes arising from floral gloss provided a dynamic visual display that could attract pollinators (Gaisterer et al., 1999). However, gloss is only expected to be effective as an attractant if the leaves are not also glossy, and, in the case of Ranunculus, not moving. Glossiness is hard to assess in a reproducible manner, and therefore hard to compare between leaves and petals, although recent attempts have been made (Whitney et al., 2012; Papiorek et al., 2014). There are inherent limitations to measuring gloss, such as the multidimensional nature of gloss perception and the fact that glossiness varies hugely with ambient light conditions (Chadwick & Kentridge, 2015). In addition, the amount of gloss does not depend solely on the geometry of the epidermis. For instance, micropatterning of the cell surface can greatly diminish the reflectivity of the surface while the anatomy of the underlying cell layers can, by contrast, enhance the optical effect (Vignolini et al., 2012a; Papiorek et al., 2014).

Structural colour

The physical features of the petal can not only modify the appearance of pigments but also generate colour directly. These ‘structural colours’ are produced by regular microscopic structures positioned just on or below the surface, which generate colour by diffraction and interference phenomena. The appearance of structural colours often changes with the observation angle, an effect known as ‘iridescence’.

Structural colouration was first reported in the leaves of a handful of plant species (Lee & Lowry, 1975) and more recently described in fruits and flowers (Fig. 1d) (Whitney et al., 2009b; Vignolini et al., 2012b; Kolle et al., 2013; Vignolini et al., 2015a). So far, all described examples of structural colours in petals are produced by pseudo-regular striations of the cuticle, acting as disordered diffraction gratings (Whitney et al., 2009b). Like gloss, iridescence is challenging to measure (van der Kooi et al., 2014; Vignolini et al., 2015a,b) and floral iridescence is not common. However, it is a widespread trait, as iridescent species can be identified in all major groups of angiosperms (Vignolini et al., 2015a,b).

We know that bumblebees, like other animals tested, can perceive iridescence (Whitney et al., 2009b) and they can also use it as a cue to detect flowers more efficiently (Whitney et al., 2016). However, like many biological photonic structures, floral diffraction gratings are not perfect as the spacing, height and width of the cuticle striations vary across the surface. Whether this disorder is biologically significant remains to be established, but recently Whitney et al. (2016) demonstrated that, unlike a perfectly regular grating, ‘imperfect’ petal iridescence enhances flower detectability without compromising correct identification. Thus, the disorder observed could have adaptive value.

Polarization

Insects use sensitivity to polarized light for a range of purposes, from general navigation to detection of food sources or nesting sites (Kraft et al., 2011; Wehner, 2014). Intriguingly, the flowers of angiosperms often display polarization patterns that probably emerge from differences in petal surface structure (i.e. different types of epidermal macro/microgeometry next to each other on a surface) (Foster et al., 2014). Although the polarization-sensitive dorsal rim of the bumblebee eye has long been thought to function only in navigation, Foster et al. (2014) showed that bees could distinguish between two different polarization patterns and identify correctly a rewarding target. Thus, polarization could act as a cue produced by physical rather than chemical means to attract pollinators, although whether this is the case in nature remains to be established.

Finally, it is worth noting that light focusing, gloss, structural colour and polarization all emerge from the same set of physical properties (cell shape and micro/macropatterning). These different cues are probably interdependent and it will be important for future studies to consider their combined, rather than separate, effects.

Conical cells, grip and slip

Landing on a flower that is moving in the wind, and finding the correct angle and approach to access the reward, is not always simple. Many plant species have evolved strategies that take advantage of the interplay between pollinator, flower surface and gravity, to limit access to certain groups of animals. A simple way in which many flowers improve the grip and handling efficiency of their flowers is to develop conical epidermal cells on the petal (Fig. 1b). Various studies have shown that these improve foraging efficiency for bees by providing an interlocking surface for their tarsal claws, which is of particular importance when flowers have to be handled at difficult angles or in windy or slippery conditions (Kevan & Lane, 1985; Whitney et al., 2009a; Alcorn et al., 2012).

Several recent reports have shown that some flowers manipulate pollinators by selective loss of these cells. Ojeda et al. (2016) analysed petal epidermal micromorphology for plants on the Macaronesian islands, and found that, in all five independent cases of transition to bird pollination, conical epidermal cells were lost. This evolutionary change is postulated to minimize nectar robbing by insects, by making the petal surface more slippery. Similarly, Papiorek et al. (2014) analysed petal epidermal morphology on all surfaces of 58 independent species. They observed that the parts of flowers that were vulnerable to nectar robbing, often the base of the corolla tube, generally had flat epidermal cells. These authors concluded that even bee-pollinated flowers had selective distribution of petal epidermal cell types in a way that maximized the advantages to the flower of controlling the physical forces acting on foraging insects.

Triggering pollen transfer

Physical forces can be recruited by plants to aid pollen transfer, and particularly pollen placement on specific parts of the pollinator’s body. Controlling pollen placement is considered particularly important for plants that flower in the same habitat as other species that use the same pollinator, as accurate pollen placement may limit the chances of interspecific pollen transfer. A recent report demonstrated that Impatiens frithii, flowering alongside four related species on Mount Cameroon, achieves specific pollen placement on the ventral part of the head of its pollinating sunbird, as opposed to on the more usual dorsal surface (Bartos & Janecek, 2014). This clever trick is achieved by a developmental twist in the shape of the flower’s nectar spur, which curves upwards. When the bird inserts its bill to feed, the physical force it exerts causes the flower to rotate 180 degrees, placing pollen below the beak. Here, the flower is using a combination of physical force from the bird with a flexible pedicel that can tolerate rotation, to achieve optimal pollen placement.

An even cleverer method of managing pollen placement through physical force is used by species of the trigger plant genus, Stylidium. In these flowers, from a family in the Asterales, the stamens and style are fused to make a single reproductive column which is held upwards and behind the corolla (Fig. 1e). This position is achieved over several minutes by changes in volume and length of cells, driven by potassium ion transport. When an insect brushes against the flower the energy stored in the column is released and it snaps forward and down (Fig. 1f), in a matter of milliseconds, to deposit pollen on the back of the insect (Joyeux, 2011). The column is then slowly pulled back to its original position over the next few minutes. This example of the use of physical forces is similar to the mechanisms used by many carnivorous plants, but in this case has been recruited to ensure successful pollination.

Slower movements can also prove highly effective at pollen placement. The stamens of the cactus Opuntia polyacantha move inwards over the course of 2–20 s after being touched by an insect of sufficient weight (Cota-Sanchez et al., 2013). This movement may force pollinators to push through the stamens to exit the flower, increasing their pollen uptake.