Nanostructural basis of rainbow-like iridescence in common bronzewing Phaps chalcoptera feathers

Structural colors are common in nature. Generally single feathers or other integuments contain only one structural color, but those of the common bronzewing display a consistent color gradient from blue to red (462-647nm) over the proximo-distal length of individual barbs. We used optical microscopy and macroand micro-spectrophotometry to characterize this color gradient, and transmission electron microscopy to investigate the nanostructure. Combining optical modeling and experimental results, we demonstrate that the rainbow-like iridescence is caused by multilayer interference from organized arrays of melanosome rods in a keratin matrix and that the color gradient results from subtle shifts in both diameter and spacing of melanosome rods. This result illustrates tight developmental control feathers and may provide inspiration for the design of multi-colored coatings or fibers. ©2014 Optical Society of America OCIS codes: (230.4170) Multilayers; (240.0310) Thin films; (350.4238) Nanophotonics and photonic crystals; (330.1690) Color. References and links 1. S. Vignolini, P. J. Rudall, A. V. Rowland, A. Reed, E. Moyroud, R. B. Faden, J. J. Baumberg, B. J. Glover, and U. Steiner, “Pointillist structural color in Pollia fruit,” Proc. Natl. Acad. Sci. U.S.A. 109(39), 15712–15715 (2012). 2. H. M. Whitney, M. Kolle, P. Andrew, L. Chittka, U. Steiner, and B. J. Glover, “Floral iridescence, produced by diffractive optics, acts as a cue for animal pollinators,” Science 323(5910), 130–133 (2009). 3. C. Grégoire, “Topography of the organic components in mother-of pearl,” J. Biophys. Biochem. Cytol. 3(5), 797–808 (1957). 4. T. M. Jordan, J. C. Partridge, and N. W. Roberts, “Non-polarizing broadband multilayer reflectors in fish,” Nat. Photonics 6(11), 759–763 (2012). 5. M. D. Shawkey, N. I. Morehouse, and P. Vukusic, “A protean palette: colour materials and mixing in birds and butterflies,” J. R. Soc. Interface 6(Suppl 2), S221–S231 (2009). 6. D. G. Stavenga, H. L. Leertouwer, T. Hariyama, H. A. De Raedt, and B. D. Wilts, “Sexual dichromatism of the damselfly Calopteryx japonica caused by a melanin-chitin multilayer in the male wing veins,” PLoS ONE 7(11), e49743 (2012). 7. M. R. Nixon, A. G. Orr, and P. Vukusic, “Subtle design changes control the difference in colour reflection from the dorsal and ventral wing-membrane surfaces of the damselfly Matronoides cyaneipennis,” Opt. Express 21(2), 1479–1488 (2013). 8. D. G. Stavenga, B. D. Wilts, H. L. Leertouwer, and T. Hariyama, “Polarized iridescence of the multilayered elytra of the Japanese jewel beetle, Chrysochroa fulgidissima,” Philos. Trans. R. Soc. Lond. B 366(1565), 709– 723 (2011). 9. A. E. Seago, P. Brady, J.-P. Vigneron, and T. D. Schultz, “Gold bugs and beyond: a review of iridescence and structural colour mechanisms in beetles (Coleoptera),” J. R. Soc. Interface 6(Suppl 2), S165–S184 (2009). 10. H. Ghiradella, “Light and color on the wing: structural colors in butterflies and moths,” Appl. Opt. 30(24), 3492–3500 (1991). 11. S. M. Doucet, M. D. Shawkey, G. E. Hill, and R. Montgomerie, “Iridescent plumage in satin bowerbirds: structure, mechanisms and nanostructural predictors of individual variation in colour,” J. Exp. Biol. 209(2), 380–390 (2006). #208397 $15.00 USD Received 25 Mar 2014; revised 25 May 2014; accepted 28 May 2014; published 6 Jun 2014 (C) 2014 OSA 16 June 2014 | Vol. 22, No. 12 | DOI:10.1364/OE.22.014625 | OPTICS EXPRESS 14625 12. H. Yin, L. Shi, J. Sha, Y. Li, Y. Qin, B. Dong, S. Meyer, X. Liu, L. Zhao, and J. Zi, “Iridescence in the neck feathers of domestic pigeons,” Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 74(5), 051916 (2006). 13. D. G. Stavenga, H. L. Leertouwer, N. J. Marshall, and D. Osorio, “Dramatic colour changes in a bird of paradise caused by uniquely structured breast feather barbules,” Proc. Biol. Sci. 278(1715), 2098–2104 (2011). 14. C. M. Eliason and M. D. Shawkey, “A photonic heterostructure produces diverse iridescent colours in duck wing patches,” J. R. Soc. Interface 9(74), 2279–2289 (2012). 15. J. Zi, X. D. Yu, Y. Z. Li, X. H. Hu, C. Xu, X. J. Wang, X. H. Liu, and R. T. Fu, “Coloration strategies in peacock feathers,” Proc. Natl. Acad. Sci. U.S.A. 100(22), 12576–12578 (2003). 16. C. M. Eliason, P.-P. Bitton, and M. D. Shawkey, “How hollow melanosomes affect iridescent colour production in birds,” Proc. Biol. Sci. 280(1767), 20131505 (2013). 17. H. Noh, S. F. Liew, V. Saranathan, S. G. J. Mochrie, R. O. Prum, E. R. Dufresne, and H. Cao, “How Noniridescent Colors Are Generated by Quasi-ordered Structures of Bird Feathers,” Adv. Mater. 22(26-27), 2871–2880 (2010). 18. H. K. Snyder, R. Maia, L. D’Alba, A. J. Shultz, K. M. C. Rowe, K. C. Rowe, and M. D. Shawkey, “Iridescent colour production in hairs of blind golden moles (Chrysochloridae),” Biol. Lett. 8(3), 393–396 (2012). 19. G. E. Hill and K. J. McGraw, Bird coloration: function and evolution (Harvard University Press, Cambridge, MA, 2006), Vol. 2. 20. S. Kinoshita, Structural colors in the realm of nature (World Scientific, Singapore, 2008). 21. G. Hill and K. McGraw, Bird coloration: mechanisms and measurements (Harvard University Press, Cambridge, MA, 2006), Vol. 1. 22. J. Sun, B. Bhushan, and J. Tong, “Structural coloration in nature,” RSC Advances 3(35), 14862–14889 (2013). 23. M. G. Meadows, N. I. Morehouse, R. L. Rutowski, J. M. Douglas, and K. J. McGraw, “Quantifying iridescent coloration in animals: a method for improving repeatability,” Behav. Ecol. Sociobiol. 65(6), 1317–1327 (2011). 24. M. D. Shawkey, A. M. Estes, L. M. Siefferman, and G. E. Hill, “Nanostructure predicts intraspecific variation in ultraviolet-blue plumage colour,” Proc. Biol. Sci. 270(1523), 1455–1460 (2003). 25. R. Maia, C. M. Eliason, P. P. Bitton, S. M. Doucet, and M. D. Shawkey, “Pavo: An R Package for the Analysis, Visualization and Organization of Spectral Data,” Methods. Ecol. Evol. 4, 906–913 (2013). 26. R. M. Azzam and N. M. Bashara, Ellipsometry and polarized light (North-Holland, Amsterdam, Netherland, 1977). 27. G. Jellison, Jr., “Data analysis for spectroscopic ellipsometry,” Thin Solid Films 234(1-2), 416–422 (1993). 28. H. L. Leertouwer, B. D. Wilts, and D. G. Stavenga, “Refractive index and dispersion of butterfly chitin and bird keratin measured by polarizing interference microscopy,” Opt. Express 19(24), 24061–24066 (2011). 29. V. T. Kampen, “Optical properties of hair,” Master Thesis (Technical University of Eindhoven, 2009). 30. M. L. Wolbarsht, A. W. Walsh, and G. George, “Melanin, a unique biological absorber,” Appl. Opt. 20(13), 2184–2186 (1981). 31. M. F. Land, “The physics and biology of animal reflectors,” Prog. Biophys. Mol. Biol. 24, 75–106 (1972). 32. S. Yoshioka and S. Kinoshita, “Direct determination of the refractive index of natural multilayer systems,” Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 83(5), 051917 (2011). 33. A. H. Sihvola, Electromagnetic mixing formulas and applications (The Institution of Engineering and Technology United Kingdom, 1999). 34. A. Garahan, L. Pilon, J. Yin, and I. Saxena, “Effective optical properties of absorbing nanoporous and nanocomposite thin films,” J. Appl. Phys. 101(1), 014320 (2007). 35. R. O. Prum, “Development and evolutionary origin of feathers,” J. Exp. Zool. 285(4), 291–306 (1999). 36. R. Maia, R. H. F. Macedo, and M. D. Shawkey, “Nanostructural self-assembly of iridescent feather barbules through depletion attraction of melanosomes during keratinization,” J. R. Soc. Interface 9(69), 734–743 (2012). 37. D. G. Stavenga, H. L. Leertouwer, P. Pirih, and M. F. Wehling, “Imaging scatterometry of butterfly wing scales,” Opt. Express 17(1), 193–202 (2009). 38. B. D. Wilts, K. Michielsen, H. De Raedt, and D. G. Stavenga, “Sparkling feather reflections of a bird-ofparadise explained by finite-difference time-domain modeling,” Proc. Natl. Acad. Sci. U.S.A. 111(12), 4363– 4368 (2014). 39. H. Durrer and W. Villiger, “Schillerfarben der Nektarvogel (Nectarinnidae),” Rev. Suisse Zool. 69, 801–804 (1962). 40. H. Durrer and W. Villiger, “Schillerradien des Goldkuckucks (Chrysococcyx cupreus (Shaw)) im Elektronenmikroskop,” Z. Zellforsch. Mikrosk. Anat. 109(3), 407–413 (1970). 41. M. Kolle, A. Lethbridge, M. Kreysing, J. J. Baumberg, J. Aizenberg, and P. Vukusic, “Bio-Inspired Band-Gap Tunable Elastic Optical Multilayer Fibers,” Adv. Mater. 25(15), 2239–2245 (2013).

In most cases studied thus far, single feathers have only one structural color that may vary depending on the incident angle and viewing angle [20][21][22].However, the covert feathers of the common bronzewing (Phaps chalcoptera) contain a color gradient from blue to red over the proximo-distal length of individual barbs that is present even at a constant illumination and viewing angle (Fig. 1).This continuous rainbow-like iridescence is distinct from the four discontinuous colors (blue, green, yellow, brown) distributed in well-defined regions in male peacock (Pavo muticus) feathers [15], and as far as we are aware has not been described before.Moreover, it suggests a high degree of nanostructural control during feather development that may serve as inspiration for the design of multi-colored fibers or coatings.We therefore investigated the nanostructural basis of this color gradient using optical and electron microscopy, spectrophotometry and optical modeling.

Barbule macrostructure
An iridescent covert feather of a male common bronzewing (Phaps chalcoptera) was obtained from the National Museum of Natural History (Washington, D.C. USA).We used a Leica S8 APO stereomicroscope (Leica Microsystems GmbH, Wetzlar, Germany) to examine the color variation of barbules along the distal-proximal axis of individual barbs (Fig. 1)

Angle-resolved specular reflectance
We visually observed slight change in hue with observation angle in the feather (suggesting interference effects), so we quantified its iridescent properties using an AvaSpec spectrometer with a xenon light source (Avantes Inc., Broomfield, USA, beam size ~5 mm) attached to a custom-built goniometer.Before any spectral measurements were collected, two calibration steps were done to maximize reflectance and thereby ensure all measurements were done using the same protocol [23].First, we made the incident beam (  ) and detector angle (  ) equal by mounting the feather onto the stage and rotating the stage until the reflectance intensity reached a maximum.The rotation angle was  , shown in Fig. 2. Second, we tilted the stage on which the feather sat to the angle  where reflectance was maximized.Through these two adjustments, the plane of the color-producing nanostructures in the barbule was made perpendicular to the bisector line of the angle between incidence and detection directions [23].To quantify the effect of incident angle, we measured the specular reflectance at incident angles varying from 10  to 45  by 5  increments (Fig. 3).Because the red and blue regions were smaller than the beam size (5 mm) [Figs.1(a)], we only performed these measurements on the green portions.Fig. 2. Schematic for geometry of spectrometer for specular reflectance measurement. is the angle between the incident beam and the surface normal; similarly,  is the angle between the detector and the surface normal; and  is the angle between feather rachis (bold black line with an arrow) and the plane-of-incidence (yellow plane).The values of  and  can be read directly from the spectrometer.The arrow of the feather stands for the distal end of the feather. is the angle between the stage plane on which the feather lies and the horizontal plane.

Single barbule normal reflectance
To precisely quantify the color of differently colored sections of the feather, we used a CRAIC AX10 UV-Visible-NIR microspectrophotometer (MSP) (CRAIC Technologies, Inc, San Dimas, USA, beam size ~4 m  , range 300-800 nm) to measure (once for each) the normal reflectance of single barbules across the different color regions of the feather (Fig. 4).We then measured the reflectance of 10 green barbules in the same barb that was then examined with transmission electron microscopy.We then performed three measurements on the normal reflectance at p-and s-polarized illumination for barbules with different colors.

Barbule nanostructure
To investigate the nanostructures of barbules, we prepared ultra-thin cross sections following the procedure described by Shawkey et al. [24].We separated one barb containing all colors from blue to red into three different color regions (blue, green and red) with the help of an optical microscope.We again cut each color region into two parts to increase the number of pieces for sampling.We then dehydrated barbs and barbules using 100% ethanol for 20 minutes twice and sequentially infiltrated them with 15%, 50%, 70% and 100% Embed 812 resin.Each infiltration step was performed on a Thermolyne Vari-Mix rocker (Thermo Scientific, Waltham, USA) for about 24 hours.Next, we placed embedded medium and samples into block molds and cured them at 60 C  overnight.We trimmed the blocks with a Leica S6 EM-Trim 2 (Leica Microsystems GmbH, Wetzlar, Germany) and cut 80 nm thick ultrathin sections with an Ultra 45 diamond knife (Diatome Ltd, Biel, Switzerland) on a Leica UC-6 ultramicrotome (Leica Microsystems GmbH, Wetzlar, Germany).Sections were transferred onto the copper grids that were viewed under a JEM-1230 transmission electron microscope (JEOL Ltd, Tokyo, Japan).From the obtained TEM images, we measured the following parameters for the three differently colored regions 100 times via the software Image J (http://fiji.sc/Fiji): 1) diameters of melanosomes, m d ; 2) spacing (edge-to-edge distance) between melanosome layers, a ; and 3) spacing (center-to-center distance) between adjacent melanosomes in the same layer, 0 d (Table 1).Because the end-to-end spacing (35  15 nm) between melanosome rods in each melanosome layer is small compared to the length of melanosome rods (870  112 nm), we can assume infinite melanosome cylinders are evenly spaced (spacing = 0 d ) in each melanosome layer and the volume fraction of melanosomes in each layer can be calculated as

Rainbow-like iridescent reflectance
The iridescent color of the bronzewing is limited to the exposed side of the feathers, as illustrated in Fig. 1(a).The color varies from red to blue along the distal-proximal gradient.These color differences are shown more clearly in magnified images for barbules in distal, middle and proximal regions [Figs.
. The barbules are oriented uniformly.Angleresolved specular reflectance of the green region of the feather is shown in Fig. 3, with the incident angle ranging from 10  to 45  by 5  increments.As the incident angle increased, the reflectance was blue-shifted (from 531nm to 492 nm) and the intensity decreased.The range of hues (we define hue as the wavelength of peak reflectance) across the whole feather, as measured by microspectrophotometer (MSP), is 185 nm [462nm to 647 nm, Fig. 4(d)].Multilayer structures in red, green and blue barbules are similar (Fig. 12) and the size and spacing of melanosomes are slightly different.The mean diameter of melanosomes   m d in distal barbules is around 10 nm larger than that in middle and proximal barbules, which are similar to one another (Table 1).The melanosome layer spacing ( a , the keratin layer thickness) is very similar in distal and middle barbules, but is about 10 nm smaller in proximal barbules.The spacing between neighboring melanosomes in the same layer ( 0 d ) is nearly identical, resulting in little variation in volume fraction (packing density, mel V ) of melanosomes in the keratin matrix for each melanosome layer (Table 1).  1 distance between edges of neighboring melanosome layers, which is also the thickness of keratin layer 2 distance between centers of neighboring melanosomes in the same layer 3 volume fraction of melanosomes in each melanosome layer

Multilayer interference modeling
We used a standard multilayer interference model [11,26,27] to calculate the theoretical reflectance for barbules of different colors (see model details in Appendix).In this planar multilayer system (Fig. 6), the top layer is the thin layer of keratin with thickness 1 d 12 5  nm.Below it are layers of alternating melanosomes and keratin.Thickness for all melanosome layers and keratin layers in individual barbules is consistent (Table 1), therefore we assume that the thickness of melanosome layer and keratin layer are constant.Complex refractive index (RI) is described as n n i  , where the real component n is related to the phase velocity and the imaginary part  indicates the absorption of the medium.Here  is wavelength of light in the unit of nm.Keratin is almost transparent and has negligible absorption [29], thus we assume ker 0   .a , the spacing between melanosomes layers and 0 d , the spacing between neighboring melanosomes.
The RI for melanin is less well characterized due to its strong absorption [30].In previous studies [31], mel n was assumed to be 2, and some modeling results have in some cases confirmed this value [11,14] while others suggested that it is lower than 2 [6,8,32].Taking dispersion effect of RI into consideration, we refer to the empirically measured values for a dense melanin-like substance [32].
Real part: In Eqs. ( 2) and (3),  is wavelength of light in nm.
Because the melanosome layer is a mixture of melanosomes and keratin [Fig.5(a) and Table 1], the RI for this layer can be corrected by averaging the refractive index of two components based on various effective medium theories [33].In this case, the volume averaging theory, applicable for absorbing materials is used to calculate the average RI of melanosome layer [34]: The constants A and B can be determined using the following equations: ker ker 2 2(1 ) In Eqs. ( 7) and ( 8), mel V is the average melanosome volume fraction in each melanosome layer (Table 1  The mean values for melanosome layer thickness, keratin layer thickness, and melanosome inner-layer volume fraction (Table 1) for green barbules were used in modeling multilayer interference.The model result was compared with average measured spectra of green barbules on the same barb for TEM imaging (Section 2.2.2).The measured peak intensity is relative parameter which is dependent on the white reference, so we standardize all the intensities in measured and modeled spectra to 1, as has been done in other literature [16].Figure 7(a) shows that the modeled peak wavelength, peak width and peak shape match closely with those in the experimental curve.Since there is still some color variation in the same color region (e.g.620~650 nm in red region), the average dimensional values from TEM images include both (e.g.red) colors with shorter wavelength and longer wavelength.To further test the contribution of multilayer structure to the color gradient in the feather, we obtained the theoretical color range (hues for bluest and reddest barbules) by modeling the color resulting from the smallest layer thickness for blue barbules (75.2 nm for melanosome layer and 63.4 nm for keratin layer) and the largest layer thickness for red barbules (99.2 nm for melanosome layer and 100.3 nm for keratin layer).The predicted hues are 477 nm and 638 nm for the bluest and reddest barbules, respectively, Fig. 7  The modeling calculations for reflectance as a function of angle predict that hue should decrease with increasing incident angle, as is observed in our empirical results (Fig. 8).Taking into consideration the standard deviation of thickness of melanin and keratin layers in green barbules, we obtain the color range where the measured points for green barbules fall (Fig. 13).Additionally, s-and p-polarized specular reflectance spectra at normal incident illumination almost overlap each other and have an identical wavelength of peak reflectance with unpolarized incidence light for all colored barbules and we only show polarized spectra for the green barbule (Fig. 9).These results are consistent with TEM images (Fig. 5) where we have observed the periodicity perpendicular to the thickness direction and not within the melanosome layers (parallel to the thickness direction).This is also consistent with polarization dependent iridescence in negative controls with two dimensional photonic crystal structure in green-winged teal (Anas.carolinensis) [14] and peacock (Pavo cristatus) [15].Based on this model, we explored how variation in the nanostructure could potentially affect color.First, we found that small changes (  12 nm) in the thickness of keratin cortex (outmost layer) result in only  7 nm variation in the value of reflectance maximum (Fig. 10).Thus, the outer cortex plays a relatively small role in the color production of the common bronzewing.Experimentally, it was difficult to precisely measure the thickness of outmost layer because of the possible shrinkage of the barbule boundaries during embedding.The result of our model, however, demonstrates that a precise value for the thickness of the keratin cortex layer is not critical to accurate modeling.By contrast, the number of melanosome/keratin layers can strongly affect reflected color.After six layers, the reflectance hue changes little with increasing number of layers.The full width at half maximum of reflectance peak decreases sharply and its decrement rate (negative value) with respect to increasing layers plateaus off at around 12 layers (Fig. 11).It is well know that the peak reflectance intensity increases with the number of layers for a multilayer system [20] but the increment of intensity for bronzewing feathers declines after 5 layers.The fact that increment rate of intensity is less than 1% after 12 layers indicates that there is diminishing reward for increasing the number of layers beyond 12. Therefore a sharp peak with enough reflectance intensity can be produced with about 12 layers of melanin and keratin (~6 melanosome layers), the number most frequently found in rainbow-like iridescent bronzewing feathers.Increment Rate of Intensity / % Fig. 11. Green solid line is the decrement rate of full width at half maximum (FWHM), which is the derivative of FWHM with respect to layer number and blue dash line is the increment rate of intensity which is the derivative of intensity with respect to layer number.

Conclusion
We report here for the first time a continuous color span ranging from blue to red in a single feather.We explain by optical modeling and TEM results that this large span (462-647 nm) is caused by subtle shifts in both spacing and diameter of melanosomes in a multilayer structure.Although spacing differences in multilayers are common, subtle variation in melanosome size has never before been reported.This is particularly intriguing because it necessitates the existence of a mechanism in the developing feather cells to precisely control (within a few nanometers) the size of the melanosomes.The diameters of melanosomes increase by 10 nm from blue (proximal) to red (distal) barbules and we hypothesize that this spatial size control is regulated by decreasing size of melanosomes as feather development proceeds, since the distal barbules grow earlier than proximal ones [35].How such fine control over size and spacing along the barb is achieved is an intriguing question that has barely been addressed.Maia et al. [36] proposed that depletion attraction of melanosomes drives arrangement of melanosomes to aggregate into a single one-layer film beneath the barbule surface during keratinization.However, the more complex and precise structuring observed here suggests additional forces may also be important in self-assembly of melanosomes.Whether these forces are subject to perturbation by external stressors, for example, will suggest whether they may serve as honest indicators of quality, as has been suggested for other iridescent feathers [36].Such multilayers of alternating high and low refractive index layers have also been utilized by many others animals, ranging from nacres of mollusks [3], fish scales [4], beetle elytra [8], butterfly wing scales [37] to damselfly wing veins [6,7].However, use of multilayer solid melanosome rods in a keratin matrix for color production has only been reported in the bird of paradise (Parotia lawesii) [13,38] since seminal early studies [39,40].Therefore, this paper may provide new inspiration for the design of multi-colored coatings or fibers from multilayer of high refractive index nanoparticles in a low refractive index matrix, which may be used for antireflection or spectral filtering.For example, bio-inspired tunable colored fibers have been made via multilayer rolling [41], but these are only a single color.Our results suggest that color gradients can be produced using spaced layers of high refractive index particles (e.g.melanosome particles) whose size can be tuned to control the thickness of these layers and whose volume fraction can be manipulated to change the refractive index of these layers, providing a novel strategy for the design of synthetic multilayer structures with optical properties.

Appendix
The matrix method for multilayer interference calculation we used was developed by Azzam and Bashara [26] and has been used in both material science [27] and biological field [11].First, we need to define three matrices as follows: Interface matrix for j th interface In Eq. ( 9), j r and j t are Fresnel reflection and transmission coefficient.For p-polorization The scattering matrix including the reflection and transmission properties of multilayer structure In Eq. ( 16), r is the reflection coefficient which can be calculated from the scattering matrix:

Fig. 1 .
Fig. 1. a) Stereo light microscope image.Panels b-d) show higher magnification images of barbules in distal region, middle region, and proximal regions, respectively.Scale bars a) 2 mm, b-d) 200 m  .

Fig. 3 .
Fig. 3. Specular reflectance spectra for the green region of barbules at incident angles from 10  to 45  .

Fig. 4 .
Fig. 4. Panels a-c) microscopic images of blue, green and red barbules from MSP, respectively and small black squares in the middle indicate the sampled spot (the length of the black square is 4 m ).d) Color variation of the barbules measured by MSP, the vertical axis is the normalized reflectance intensity with arbitrary unit.The color for each curve is based on the human visual perception according to the standard CIE1931[25].

Fig. 5 .
Fig. 5. TEM images of a) cross section of a red barbule, b) longitudinal section of a green barbule and c) cross section of a non-iridescent brown barbule.Scale bars: a-c) 500 nm.

Table 1 .
Spacing and diameter of melanosomes in the barbule nanostructure measured using TEM results (Errors are standard deviation from 100 measurements)

Fig. 7 .
Fig. 7. a) Measured (green line) and modeled spectra (black line) on the average layer thickness for barbules in green region.b-c) The modeled bluest and reddest spectra (black lines) based on largest and smallest thickness of melanosome and keratin layers in blue and red barbules; and the blue and red curves are the bluest and reddest spectra measured by MSP, respectively.

Fig. 8 .
Fig.8.Modeled results for angle-resolved reflectance spectrum for middle barbules (green solid line).The black square data points are obtained from experimental spectra in Fig.3.

Fig. 9 .
Fig. 9. Reflectance of green barbule measured by MSP using differently polarized input beams.The blue curve is the unpolarized reflectance, and the red, green are the s-and ppolarization, respectively.

Fig. 10 .
Fig. 10.The hue dependence on the outmost keratin layer (cortex layer) thickness based on multilayer modeling result.

#Fig. 13 .
Fig. 12. TEM images of cross-sections of barbules with different colors under the same magnification: a) red, b) green and c) blue.Scale bars 100nm