A trick of the light? Convergent evolution of disordered flower surfaces and the blue light they scatter.

In a new paper in Nature, we explore the role that nanoscale surface sculpturing of the petal plays in optical properties of flowers and their interactions with pollinators. We report that fine scale ridges that generate scattered light in the blue-UV parts of the spectrum have evolved convergently in most of the main groups of flowering plants.

Like Comment

The paper in Nature is here: http://go.nature.com/2impD10

Our work with structural colour first started with the observation by a former post-doc in my lab, Heather Whitney, that the dark red pigmented patch on the petal of Hibiscus trionum looked different colours when the sun shone – producing an iridescent sheen on top of the red pigment. In 2009 we published a paper1 showing that the petal surface of this flower has fine ridges of cuticle on top of the epidermal cells, which act like a diffraction grating – generating reflection of different colours depending on the angle from which the flower is observed.

But we were intrigued to know whether this was a unique phenomenon, or whether other flowers could also play the same trick with the light. So, in 2011, we started to use the fantastic living collections of plants at the Cambridge University Botanic Garden and the Royal Botanic Gardens Kew to explore how widespread structurally coloured flowers are. We used a targeted approach – knowing that the cells underlying the cuticle have to be approximately flat to generate an angle-dependent colour effect, we sampled in those angiosperm families where petals or tepals are known to frequently have flat epidermal cells.

After spending time in the garden with our physicist colleagues and taking samples back to the lab to measure them, we identified flowers with nanoscale cuticular ridges that generate a measurable colour signal in almost all branches of the angiosperm phylogenetic tree – in families from the monocots, the early diverging eudicot order Ranunculales, both major rosid groups and both major asterid groups. Structural colour in flowers is not common, but it is widespread.

However, while measuring the optical properties of these flowers, we discovered something surprising. The iridescent signal they all produce is quite weak, as a result of the disorder in the surface structure. Instead of a perfectly regular diffraction grating, the cuticle ridges slightly vary in height, width and spacing, within each individual flower. What was surprising, was that all the different flowers we analysed produce a very similar optical signal –they scatter light in the blue-UV part of the spectrum, over a specific range of angles. We call this effect “the blue halo”.

Given that all of these different flowers had convergently evolved surface structures with just the right degree of disorder to produce the blue halo, we hypothesised that perhaps it was this, rather than the weak iridescence we had first noticed, which was being selected for. We then decided to fabricate artificial surfaces which replicate the “blue halo effect”. When we made artificial flowers with these surfaces we found that bees could easily see them, even on top of pigment backgrounds (such as yellow) which make them invisible to the human eye. This was exciting, because many of our blue halo flowers do not look blue to us – but it appears that they do to a bee! We then discovered that the blue halo had a significant effect on foraging efficiency, shortening the time it took a bee to find flowers that do not contain a blue pigment and speeding up their rate of foraging.

True blue is a difficult colour for plants to produce chemically: it requires the enzymatic capability to add extra hydroxyl groups to the anthocyanin molecule as well as modifications of vacuolar pH, and generations of breeders have worked to try to engineer blue shades in popular flowers like roses and chrysanthemum. Given that insect vision is shifted towards the blue-UV relative to human vision, we conclude that the blue halo has evolved convergently as an alternative way for flowers to produce an optical signal that is highly visible to insect pollinators. The next challenge for us is to understand the developmental biology of the system – just how do flowers produce such a regular (but sufficiently irregular) pattern of cuticle ridges on their surfaces? – but this is another story!


1 Whitney, H. et al. Floral iridescence, produced by diffractive optics, acts as cue for animal pollinators. Science 323, 130-133 (2009).

The paper in Nature is here:  http://rdcu.be/wSd3

Beverley Glover

Director of Cambridge University Botanic Garden, University of Cambridge