A meta-analysis of butterfly structural colors: their color range, distribution, and biological production

Butterfly scales are among the richest natural sources of optical nanostructures, which produce structural color and iridescence. Several recurring nanostructure types have been described, such as ridge multilayers, gyroids, and lower lamina thin films. While the optical mechanisms of these nanostructure classes are known, their phylogenetic distributions and functional ranges have not been described in detail. In this Review, we examine a century of research on the biological production of structural colors, including their evolution, development, and genetic regulation. We also create a database of more than 300 optical nanostructures in butterflies and conduct a meta-analysis of the color range, abundance, and phylogenetic distribution of each nanostructure class. Butterfly structural colors are ubiquitous in short wavelengths but extremely rare in long wavelengths, especially red. In particular, blue wavelengths (around 450 nm) occur in more clades and are produced by more kinds of nanostructures than other hues. Nanostructure categories differ in prevalence, phylogenetic distribution, color range, and brightness. For example, lamina thin films are the least bright; perforated lumen multilayers occur most often but are almost entirely restricted to the family Lycaenidae; and 3D photonic crystals, including gyroids, have the narrowest wavelength range (from about 450 to 550 nm). We discuss the implications of these patterns in terms of nanostructure evolution, physical constraint, and relationships to pigmentary color. Finally, we highlight opportunities for future research, such as analyses of subadult and Hesperid structural colors and the identification of genes that directly build the nanostructures, with relevance for biomimetic engineering.

Methods
Several intersecting approaches were used to search for articles, book chapters, and theses reporting nanostructures that produce structural colors in butterflies. First, Google Scholar searches were run with combinations of these keywords: structural color, butterfly, Lepidoptera, iridescence, scale, and Ghiradella. Database searches brought up tens of thousands of hits, many of which were not pertinent, so we used high-quality results to find cited and citing references and noted every species that was mentioned in connection with structural color, iridescence, or derived scale morphology. Finally, a database search was run on each species name that had been mentioned in any prior included reference. When a search on species name returned many results, it was searched again in combination with the keywords. If the search on species name returned no relevant results, we tried searches with only the genus name and checked for alternate nomenclature. We continued snowballing references and running database searches on species until we could no longer find any new taxa mentioned in connection with structural color. Criteria for inclusion were that the article must include either (1) reflectance measurements or (2) electron microscope images of non-transparent (i.e. colored) optical nanostructures in a butterfly. We also included studies which provided additional characterizations (e.g. absorption measurements, mathematical modeling, scatterometry) for structures that had been included on the basis of (1) or (2).

This literature review strategy yielded 187 included references which described 421 potential structures from 378 species, all of which are included in this dataset. Before further analysis, we secondarily excluded entries that were presented as non-photonic comparisons to structurally colored specimens. Only seven reported structures occurred outside adult wing scales, which were all multilayer broadband reflectors in the pupal cuticle (Neville, 1977; Steinbrecht, 1985; Steinbrecht et al., 1985). We therefore narrowed our focus to structures located in scales or bristles on the adult, which can be homologized and directly compared in subsequent analyses. After filtering, there was pairwise complete data on both color and morphology for 314 optical nanostructures from 287 species. Some species had multiple structural colors on different body parts (e.g. blue dorsal and green ventral wings in Cyanophrys remus and Albulina metallica; Biró et al., 2007). 

To compare color between structures, we recorded the peak reflected wavelength (i.e. hue) and the percent reflectance at that wavelength (i.e. brightness) for each structural color. Due to iridescence, quantification of structural color is extremely sensitive to the measurement protocol, specifically illumination and detection angles, light source, reference sample, and spot size (Meadows et al., 2011). Spectroscopy methodologies were variable among the included studies making comparisons imperfect; nevertheless, the data are useful to show broad trends. When multiple spectra were available, we used the following rules for consistency. When reflectance was reported from more than one angle, the peak wavelength at the maximally reflective angle was used. If comparable reflectance data was reported from more than one study or from replicated specimens, we took their average. When reflectance data was found for both an isolated scale and the intact wing, both values were noted, but the intact wing reflectance was preferentially used in comparative analyses for consistency, because single-scale reflectance measures were uncommon. In cases where structures produced two reflectance peaks – as in Chrysozephyrus species with both a UV and a green peak (Imafuku, Hirose and Takeuchi, 2002) – the brighter peak was used in graphical summaries, but both were listed in the spreadsheet. Peak wavelengths were typically estimated by eye from graphs, which limited precision to a 5–10 nm window around the measured peak. This precision limit is similar to the magnitude of inter-individual variation (Imafuku, Gotoh and Takeuchi, 2002; Bálint et al., 2008). We also recorded percent reflectance at the maximally reflective wavelength (i.e. spectral intensity or ‘brightness’). When no reflectance spectra were available but a color image or a qualitative color descriptor (e.g. ‘blue’, ‘UV’) was given, the qualitative descriptor was recorded. Broadband reflectors have a similar reflectance intensity across many wavelengths, so the maximally reflecting wavelength is not a good summary of the reflector’s properties and may not be identifiable from a visual inspection of a graph. Therefore, for broadband reflectors, we only recorded a qualitative descriptor, such as ‘white’, ‘silver’, or ‘gold’. Additionally, some reddish lamina thin films that reflected in both violet and red, without a peak wavelength in either region, were handled as qualitatively ‘magenta’. Note that many reflectance spectra were likely influenced by co-occurring pigments as well as the nanostructures. 

To compare scale morphological modifications, we noted which kind of optical nanostructure was present. Generally, we followed the author’s conclusion as to which scale component caused the optical properties. If the author’s description was brief but a micrograph was provided, we assigned the structure to the same category as the well-studied examples that it most resembled. In a few cases when the proposed mechanism seemed questionable, we noted the explanation in the spreadsheet but dropped that structure from comparative analysis (for example, the proposed nanostructure was not present in the provided micrograph, or the mechanism was disputed across studies). Filled-in windows and crossrib air columns likely involve modifications to both the crossribs and microribs, and reflectance in these scales also requires the lower lamina; for simplicity, we have summarized them as crossrib bilayer structures.

Data was processed using R.