Diverse forms of nanoscale architecture generate structural colour and perform signalling functions within and between species. Structural colour is the result of the interference of light from approximately regular periodic structures; some structural disorder is, however, inevitable in biological organisms. Is this disorder functional and subject to evolutionary selection, or is it simply an unavoidable outcome of biological developmental processes? Here we show that disordered nanostructures enable flowers to produce visual signals that are salient to bees. These disordered nanostructures (identified in most major lineages of angiosperms) have distinct anatomies but convergent optical properties; they all produce angle-dependent scattered light, predominantly at short wavelengths (ultraviolet and blue). We manufactured artificial flowers with nanoscale structures that possessed tailored levels of disorder in order to investigate how foraging bumblebees respond to this optical effect. We conclude that floral nanostructures have evolved, on multiple independent occasions, an effective degree of relative spatial disorder that generates a photonic signature that is highly salient to insect pollinators.
Subscribe to Journal
Get full journal access for 1 year
only $3.90 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Vukusic, P. & Sambles, J. R. Photonic structures in biology. Nature 424, 842–845 (2003)
Whitney, H. M. et al. Floral iridescence, produced by diffractive optics, acts as a cue for animal pollinators. Science 323, 130–133 (2009)
Vignolini, S. et al. The flower of Hibiscus trionum is both visibly and measurably iridescent. New Phytol. 205, 97–101 (2015)
Whitney, H. M., Reed, A., Rands, S. A., Chittka, L. & Glover, B. J. Flower iridescence increases object detection in the insect visual system without compromising object identity. Curr. Biol. 26, 802–808 (2016)
Prum, R. O. & Torres, R. H. A Fourier tool for the analysis of coherent light scattering by bio-optical nanostructures. Integr. Comp. Biol. 43, 591–602 (2003)
Scotland, R. W. What is parallelism? Evol. Dev. 13, 214–227 (2011)
The Angiosperm Phylogeny Group III. An update of the Angiosperm Phylogeny Group Classification for the orders and families of flowering plants: APG III. Bot. J. Linn. Soc. 161, 105–121 (2009)
Endress, P. K. The evolution of floral biology in basal angiosperms. Phil. Trans. R. Soc. Lond. B 365, 411–421 (2010)
Grimaldi, D. The co-radiations of pollinating insects and angiosperms in the Cretaceous. Ann. Mo. Bot. Gard. 86, 373–406 (1999)
Grimaldi, D. & Engel, M. S. Evolution of the Insects (Cambridge Univ. Press, 2005)
Labandeira, C. C. The pollination of Mid Mesozoic seed plants and the early history of long-proboscid insects. Ann. Mo. Bot. Gard. 97, 469–513 (2010)
Vignolini, S. et al. Directional scattering from the glossy flower of Ranunculus: how the buttercup lights up your chin. J. R. Soc. Interface 9, 1295–1301 (2012)
Noda, K., Glover, B. J., Linstead, P. & Martin, C. Flower colour intensity depends on specialized cell shape controlled by a Myb-related transcription factor. Nature 369, 661–664 (1994)
Gorton, H. L. & Vogelmann, T. C. Effects of epidermal cell shape and pigmentation on optical properties of antirrhinum petals at visible and ultraviolet wavelengths. Plant Physiol. 112, 879–888 (1996)
Vignolini, S. et al. The mirror crack’d: both pigment and structure contribute to the glossy blue appearance of the mirror orchid, Ophrys speculum. New Phytol. 196, 1038–1047 (2012)
van der Kooi, C. J ., Elzenga, J. T. M ., Staal, M. & Stavenga, D. G. How to colour a flower: on the optical principles of flower coloration. Proc. R. Soc. Lond. B 283, 20160429 (2016)
van der Kooi, C. J. et al. Iridescent flowers? Contribution of surface structures to optical signaling. New Phytol. 203, 667–673 (2014)
Johansen, V. E. et al. Designing visual appearance using a structured surface. Optica 2, 239–245 (2015)
Schauer, S., Worgull, M. & Hölscher, H. Bio-inspired hierarchical micro- and nano-wrinkles obtained via mechanically directed self-assembly on shape-memory polymers. Soft Matter 13, 4328–4334 (2017)
Schneider, N. et al. Nanothermoforming of hierarchical optical components utilizing shape memory polymers as active molds. Opt. Mater. Express 4, 1895–1902 (2014)
Spaethe, J., Tautz, J. & Chittka, L. Visual constraints in foraging bumblebees: flower size and color affect search time and flight behavior. Proc. Natl Acad. Sci. USA 98, 3898–3903 (2001)
Giurfa, M., Nunez, J., Chittka, L. & Menzel, R. Colour preferences of flower-naive honeybees. J. Comp. Physiol. A 177, 247–259 (1995)
Raine, N. E., Ings, T. C., Dornhaus, A., Saleh, N. & Chittka, L. Adaptation, genetic drift, pleiotropy, and history in the evolution of bee foraging behavior. Adv. Stud. Behav. 36, 305–354 (2006)
Raine, N. E. & Chittka, L. Nectar production rates of 75 bumblebee-visited flower species in a German flora (Hymenoptera: Apidae: Bombus terrestris) . Entomol. Gen. 30, 191–192 (2007)
Raine, N. E. & Chittka, L. The adaptive significance of sensory bias in a foraging context: floral colour preferences in the bumblebee Bombus terrestris. PLoS One 2, e556 (2007)
Yoshida, K., Kondo, T., Okazaki, Y. & Katou, K. Cause of blue petal colour. Nature 373, 291 (1995)
Yoshida, K., Mori, M. & Kondo, T. Blue flower color development by anthocyanins: from chemical structure to cell physiology. Nat. Prod. Rep. 26, 884–915 (2009)
Holton, T. A. & Tanaka, Y. Blue roses — a pigment of our imagination? Trends Biotechnol. 12, 40–42 (1994)
Hondo, T. et al. Structural basis of blue-colour development in flower petals from Commelina communis. Nature 358, 515–518 (1992)
Katsumoto, Y. et al. Engineering of the rose flavonoid biosynthetic pathway successfully generated blue-hued flowers accumulating delphinidin. Plant Cell Physiol. 48, 1589–1600 (2007)
Shiono, M., Matsugaki, N. & Takeda, K. Phytochemistry: structure of the blue cornflower pigment. Nature 436, 791 (2005)
Sasaki, N. & Nakayama, T. Achievements and perspectives in biochemistry concerning anthocyanin modification for blue flower coloration. Plant Cell Physiol. 56, 28–40 (2015)
Yasin, S., Hasko, D. G. & Ahmed, H. Fabrication of <5 nm width lines in poly(methylmethacrylate) resist using a water:isopropyl alcohol developer and ultrasonically-assisted development. Appl. Phys. Lett. 78, 2760–2762 (2001)
Williams, S. S. et al. High-resolution PFPE-based molding techniques for nanofabrication of high-pattern density, sub-20 nm features: a fundamental materials approach. Nano Lett. 10, 1421–1428 (2010)
Dyer, A. G. & Chittka, L. Biological significance of distinguishing between similar colours in spectrally variable illumination: bumblebees (Bombus terrestris) as a case study. J. Comp. Physiol. A Neuroethol. Sens. Neural Behav. Physiol. 190, 105–114 (2004)
Foster, J. J. et al. Bumblebees learn polarization patterns. Curr. Biol. 24, 1415–1420 (2014)
Warton, D. I. & Hui, F. K. C. The arcsine is asinine: the analysis of proportions in ecology. Ecology 92, 3–10 (2011)
We thank M. Dorling for plant and bee care; P. Cunha for advice on e-beam lithography; and B. Wilts, J. Baumberg, R. Bateman, N. Cunniffe, N. Walker-Hale, L. Chittka, H. Whitney and M. Kolle for discussion. We acknowledge the collections at Cambridge University Botanic Garden and the Royal Botanic Gardens, Kew. This work was funded by the Leverhulme Trust (F/09741/G to B.J.G. and U.S.), BBSRC (DTG studentship to A.R. and the David Phillips fellowship (BB/K014617/1) (76933) to S.V.), the European Research Council ((ERC-2014-STG H2020 639088) to S.V.), the Herchel Smith fund (to E.M.), EU Marie Curie actions (NanoPetals to E.M. and B.J.G.), EPSRC (EP/G037221/1 to R.M.), the Winton Fund for the Physics of Sustainability and the Cambridge Trust CHESS (to T.W.), the Adolphe Merkle Foundation and the Swiss National Science Foundation (National Center of Competence in Research Bio-Inspired Materials) (U.S.). We thank the EU for funding under Marie Curie Actions I.T.N. PlaMatSu (722842) to U.S., S.V. and B.J.G.
The authors declare no competing financial interests.
Reviewer Information Nature thanks D. Deheyn and the other anonymous reviewer(s) for their contribution to the peer review of this work.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data figures and tables
a–c, Individual reflectance spectra of the three flowers in Fig. 2 are shown at scattering angles of 10°, 20° and 30°, recorded in parallel and perpendicular planes (with respect to the direction of surface striations), relative to the reflection of a white standard. d–f, Reflectance difference after subtracting parallel from perpendicular orientation measurements, revealing enhanced scattering for shorter wavelengths. a, d, A. aestivalis at negative angles; b, e, O. stricta reflectance reduced by a factor of ten to fit plot limits; c, f, H. trionum at negative angles.
Extended Data Figure 2 Optical response and anatomical parameters of petals with cuticular striations (P. mascula, L. purpurea, P. barrettiae, Tulipa ‘Queen of the Night’ and L. aureus).
a–e, Scattering measurements from flowers with differing degrees of disorder; P. mascula (a), L. purpurea (b), P. barrettiae (c), Tulipa ‘Queen of the Night’ (d) and L. aureus (e). The images are two-dimensional plots showing the spectra of the scattered light in function of the sine-scale collection angle, relative to the location of the specular reflection (set to zero degrees). The intensity of the light is represented on a blue–yellow colour scale. The left column shows the angle distribution of the scattered light in the plane perpendicular to the striation direction, whereas in the right column the direction of the striations is parallel to the plane of collection. The angle of light incidence is 45° for all measurements. The colour scale in the scattering plots (normalized to a white diffuser standard) is kept constant for the same flower, but varies from flower to flower. To emphasize the blue halo, the spectral region between 600 and 700 nm (which is not perceived by the bees) is partially masked; horizontal dotted lines with downward arrows indicate the enhanced spectral region of the blue halo. f–j, Angle-dependent spectral response mediated on the three bee photoreceptors. Each of the points in the graph corresponds to an integral over the measured spectrum at the corresponding collection angle, after it has been weighted by the normalized sensitivity curves of the three types of photoreceptors in bee eyes, ultraviolet (in violet), blue (in blue) and green (in green). The spectra in the planes perpendicular and parallel to the striation direction are reported as darker and lighter colours, respectively. See Methods for additional details. k–o, Histograms of the measured striation parameters in terms of spacing and size (height and width), extracted from the TEM images, quantifying the amount of disorder.
Extended Data Figure 3 Optical response and anatomical parameters of petals with cuticular striations (U. speciosa, G. humifusum, O. stricta, A. aestivalis and M. lindleyii).
a–c, j, k, Scattering measurements from flowers with differing degrees of disorder; U. speciosa (a), G. humifusum (b), O. stricta (c), A. aestivalis (j) and M. lindleyii (k). The images are two-dimensional plots showing the spectra of the scattered light in function of the collection angle, relative to the location of the specular reflection (set to zero degrees). The intensity of the light is represented on a blue–yellow colour scale. The upper row for each species shows the angular distribution of the scattered light in the plane perpendicular to the striation direction, whereas in the lower row the direction of the striations is parallel to the plane of collection for a range of angles. For each flower, three sets of measurements are shown for the angles of light incidence onto the petals as reported in the figure (5°, 30° and 45°). The colour scale in the scattering plots is kept constant for each pair of sample orientations, but its maximum value varies from flower to flower and between angles of incidence: a, 0.6, 0.6, 0.8; b, 1.6, 2.0, 2.0; c, 1.3, 1.5, 1.8; j, 0.35, 0.4, 0.55; k, 0.68, 0.8, 1.2 (all normalized to a white diffuser standard). To emphasize the blue halo, the spectral region between 600 and 700 nm (which is not perceived by the bees) is partially masked; horizontal dotted lines indicate the enhanced spectral region of the blue halo. d, f, h, l, n, The top plot on the right for each species reports the angle-dependent spectral response mediated by the three bee photoreceptors. Each of the points in the graph corresponds to an integral over the measured spectrum at the corresponding collection angle, after it has been weighted by the normalized sensitivity curves of the three types of photoreceptors in bee eyes, ultraviolet (in violet), blue (in blue) and green (in green). The spectra in the planes perpendicular and parallel to the striation direction are reported as darker and lighter colours, respectively. See Methods for additional details. e, g, i, m, o, Histograms of the measured striation parameters in terms of spacing and size (height and width), extracted from the TEM images, quantifying the amount of disorder.
Additional scattering measurements for flowers in which the effect is particularly hard to identify from the single batch of measurements presented in Fig. 2 and Extended Data Figs 2, 3. Datasets were collected separately for different years to analyse the effect multiple times for the same species. Some of the measurements show the effect of the halo better than others; for consistency, the main text presents only a single dataset, in which all the flowers had been measured recently with the same experimental setup. This figure introduces samples from additional datasets, with minor differences (mainly in terms of the intensity of the lamp and the angular resolution) between the experimental setups. a–d, Scattering measurements from flowers with differing degrees of disorder; A. aestivalis (a), M. lindleyii (b), P. mascula (c) and U. speciosa (d). The images are two-dimensional plots showing the spectra of the scattered light in function of the sine-scale collection angle, relative to the location of the specular reflection (set to zero degrees). The left column shows the angle distribution of the scattered light in the plane perpendicular to the striation direction, whereas in the right column the direction of the striations is parallel to the plane of collection. The angle of light incidence is 45° for all measurements. The colour scale in the scattering plots is kept constant for the same flower, but it varies from flower to flower. To emphasize the blue halo, the spectral region between 600 and 700 nm (which is not perceived by the bees) is partially masked; horizontal dotted lines with downward arrows indicate the enhanced spectral region of the blue halo. e–h, Angle-dependent spectral response mediated on the three bee photoreceptors. See Methods for additional details.
FDTD simulation of the scattering responses of rectangular gratings with parameters and disorder according to the measured flower parameters of A. aestivalis (a), G. humifusum (b), H. trionum (c), L. aureus (d), L. purpurea (e), M. lindleyii (f), O. stricta (g), P. mascula (h), P. barrettiae (i), Tulipa ‘Queen of the Night’(j) and U. speciosa (k). The intensity of the light is represented on a blue–yellow colour scale. The bands denoted by stars (a–c, f), containing the zero-order reflections, were reduced in intensity by a factor of three compared to the other regions of the graph. The additional graphs (l–s, grey box) demonstrate the process of averaging individual results to reveal the representative scattering pattern associated with the corresponding amount of disorder. These graphs contain the FDTD simulation results of rectangular gratings with dimensions equivalent to the H. trionum parameters, as in Fig. 3 (730 nm height, 730 nm width, 1,300 nm spacing; standard deviations: 0.27 height, 0.16 width and 0.29 spacing). l–o, Scattering plots of individual FDTD simulation results (as described in Methods). p, Average of scattering plots in l–o. The reduction in first-order interference and the appearance of a blue halo can be observed using a small sample number, but they become representative only when using averages of larger sample numbers, as in q (20×), r (40×) and s (60×). This observation also confirms that it is necessary to average a number of measurements taken at the same configuration or to illuminate a large area, in order to capture the colour-dependent scattering in a real flower petal with disorder. Depending on the size of the illuminated area, the measurement of a semi-disordered surface may have a similar appearance to the averages shown here, or appear randomly pixelated17. The minimum illuminated area required to observe representative distributions on flower petals is smaller than that required for the artificial samples, because natural striations vary slightly in their direction of propagation, whereas our artificial lines remained strictly parallel (providing less parameter variation for the same 2D area).
a–f, Optical measurements at 45° angle of incident light are presented for an M. lindleyii flower petal (a–c) and its peeled-off epidermal layer (d–f). For the entire petal (a, b), UV and yellow pigment colouration is visible at all angles and in both scanning directions (perpendicular and parallel to the direction of the striations). This produces high relative values for the UV receptor and green receptor, respectively (c). For colours corresponding to all three photoreceptors, more light is scattered perpendicular to the striations than parallel to them. At the same time, the overlapping optical signal of pigmentation makes it hard to recognize the colour trend of the surface-scattered light (the blue halo). When measuring the optical response of the striations on only the peeled-off epidermis, however, most pigments have been removed and the scattering collected perpendicularly to the striations (d) is caused by the structure itself. Almost no scattering is observed outside of the specular reflection when measuring parallel to the striation direction on peeled-off epidermis (e). In this case, the colouration of the halo becomes more apparent. The scattered light is enhanced in the low-wavelength (blue–UV) region and is most intense between −25° and +25° (f). i–l, To provide a qualitative comparison for the colouration effect observed on the peeled epidermis, we prepared a set of simulation results (using H. trionum parameters) in the same bee-receptor plot (grey box). Angle-dependent spectral response mediated on the three bee photoreceptors (g–i) in the simulations with H. trionum-derived parameters (j–l), with varying degrees of disorder (g, j, ordered; h, k, 1× natural disorder; i, l, 2× natural disorder). As a result of the subtlety of the halo colouration and the overlapping spectral sensitivity of bee photoreceptors, the relative intensity differences between receptor values are small and decrease with increasing disorder beyond those values found in actual flower species. To emphasize the blue halo in the scattering plots (a, b, d, e, j–l), the spectral region between 600 and 700 nm (which is not perceived by the bees) is partially masked; horizontal dotted lines with downward arrows indicate the enhanced spectral region of the blue halo.
Extended Data Figure 7 Artificial flowers used in behavioural experiments, appearance of yellow and blue pigmented test-squares and examples of flowers with a blue halo effect visible to the human eye.
a–d, The blue halo effect is best seen by the human eye on a dark pigmented background. Ursinia calendulifolia (a, b) and H. trionum (c, d) flowers present a striated epidermis at the base of their petals, which overlaps with a darkly pigmented zone. Anthocyanin pigment produces the dark purple colour and disordered striations produce the blue halo effect, visible at the base of the ray florets (b) or the proximal region of the petals (d). e, Schematic representation of artificial flower used in differential conditioning experiments (depicted with a yellow pigmented test-square) and photograph of a bee feeding on such a flower (with a black, perfectly iridescent test-square). Rewards or punishment are presented in the lid of the black Eppendorf tube. f, Yellow test-squares with a smooth surface (Sm) or overlain by a manufactured disordered structure (Di) are hardly discernible from one another, even if the observation angle varies. g, Schematic representation of artificial flower used in the foraging speed experiment (depicted with a yellow test-square) and photograph of a marked forager feeding on such a flower (with a black, perfectly iridescent test-square). h, Blue test-squares with a smooth surface or overlain by a manufactured disordered structure (as in g) appear identical to one another at some angles but at other angles they display distinct shades and intensities of blue. Image in a, b taken by H. Rice.
Extended Data Figure 8 Angular components of the photonic effect of striations and role of height levels in disordered gratings.
FDTD simulation spectra of rectangular gratings corresponding to H. trionum parameters, as in Fig. 3 (730 nm height, 730 nm width, 1,300 nm spacing). a, Specular reflection in the angular range from –2° to +2°, relative to the normal angle of incidence. b, Region of the blue halo between the specular reflection and the diffraction order peaks from –10° to +10°, excluding the region of specular reflection. This region was smaller than the full extent of the halo (which covers roughly between –20° and +20°) to avoid overlap with diffraction orders. c, Region spanning a large angular range from –40° to +40°, excluding the regions of a and b. This region contains the shorter wavelength parts of the first diffraction orders and the angles next to the high-intensity region of the blue halo. The change of the spectrum in the respective angular region is shown in a–c, with increasing disorder ranging from 0 (no disorder) through 1 (disorder corresponding to H. trionum parameters, standard deviations: 0.27 height, 0.16 width, 0.29 spacing) to 2, twice the standard deviation of natural disorder of H. trionum. d–f, Spectral response for the same three angular intervals as in a–c, together with the total amount of reflected light for one implementation of disorder each: no disorder (d), H. trionum levels of disorder (e), twice the disorder of H. trionum (f). The reduction in specular reflection can be observed in a, along with the reduction in thin film interference caused by the grating quasi-layer of intermediate refractive index. The fast rise of the blue halo for increasing disorder, and the stability of the effect for a wide range of disorder values, is shown in b. The quick decay of the first order diffraction components can be observed in c, along with the increased long-wavelength scattering response in this angular region. Note that the light reflected into this region is spread out over an angular range more than four times as large as the interval in b, reducing the scattered light intensity per viewing angle. The scattering response for incremental increases of striation disorder in e and f demonstrate the robustness of the blue halo effect with respect to the amount of disorder present in the striations. g, h, FDTD simulation results that reveal the role of height levels in disordered gratings. Simulation results are shown for rectangular gratings corresponding to the H. trionum parameters, as in Fig. 3 (730 nm height, 730 nm width, 1,300 nm spacing; standard deviations: 0.16 width and 0.29 spacing). g, Simulation without variation in height. The quasi-layer containing both air and grating teeth has an effective intermediate refractive index and results in thin film interference fringes. h, Simulation with one intermediate height level introduced at random in 40% of positions in the grating reduces thin-film interference and the colouration this causes. The artificial disordered gratings were manufactured by e-beam lithography from a thin film, which does not allow continuous variation in height.
Extended Data Figure 9 Behaviour of individual foragers during differential conditioning experiments.
a, Learning curve of five bees choosing from among six black smooth artificial flowers (three punishing and three rewarding). Empty circles, mean proportion of bees making a correct choice, for each 80 successive choices. White curve, fitted binomial logistic model with green shading showing 95% confidence intervals on the fitted response. The χ2 statistic and P value for the likelihood ratio test (assessing whether foragers can learn) are given at the bottom right of the panel. b, As in a, but showing the learning curves of each individual. The frequency of correct choice (rewarding flower) is calculated for every ten visits. None of the five foragers used in this experiment successfully managed to identify the rewarding flowers accurately even after 80 visits. c, Learning curve of ten individual bees choosing between black smooth and black disordered artificial flowers, as in Fig. 4b. The frequency of correct choice (rewarding flower) is calculated for every ten visits. d, As in c, but with bees choosing between black smooth and black ordered artificial flowers, as in Fig. 4c. e, As in c, but with ten bees choosing between yellow smooth and yellow disordered artificial flowers, as in Fig. 4d. f, As in c, but with 11 bees choosing between blue smooth and blue disordered artificial flowers, as in Fig. 4e. Source data
About this article
Cite this article
Moyroud, E., Wenzel, T., Middleton, R. et al. Disorder in convergent floral nanostructures enhances signalling to bees. Nature 550, 469–474 (2017). https://doi.org/10.1038/nature24285
Scientific Reports (2020)
New Phytologist (2020)
Scientific Reports (2020)
PeERF1, a SHINE-Like Transcription Factor, Is Involved in Nanoridge Development on Lip Epidermis of Phalaenopsis Flowers
Frontiers in Plant Science (2020)
Current Biology (2020)