The evolution of brilliant iridescence in birds

The evolution of brilliant iridescence in birds | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article The evolution of brilliant iridescence in birds Michael Nicolai, Gerben Debruyn, Rauri Bowie, Scott Edwards, Miguel Carneiro, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8020267/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Colors of iridescent feathers are some of the most striking in nature, but when and how they evolved remains elusive. Using a dataset of 71,536 body patches from 5,755 bird species, we reconstruct the evolution of iridescence. We find that brilliant iridescence, such as that of peacocks, is likely ancestral in birds, and possibly in dinosaurs, but was lost and regained multiple times over evolutionary history. Transitions from weak to brilliant iridescence were gradual, and iridescence evolved from both melanized and non-melanized patches. These findings reshape our understanding of color evolution in birds and their ancestors. Biological sciences/Evolution Biological sciences/Zoology Figures Figure 1 Figure 2 Figure 3 Figure 4 Main text Birds are among the most colorful animals, partly due to the diversity of their color production mechanisms. For example, bird colors can arise from the deposition of various pigments, such as carotenoids, melanins, and protoporphyrins ( 1 ). Additionally, they exhibit iridescent structural colors, which result from light scattering off highly organized nano-scale arrays of melanin-containing organelles called melanosomes ( 1,2 ). These nanostructures produce iridescent colors by the scattering of light from alternating layers of different refractive indices. While a single layer of melanosomes under the keratin cortex is sufficient to produce iridescence ( 3-6 ), additional variation in dimensions and numbers of layers creates additional complexity. As such, melanosomes can vary in size, spacing, and/or shape (being flattened, hollow or both), and can also be organized in layers of different sizes and arrangements, which increases the number of interfaces available for scattering ( 3,7 ) and thereby produces more intense colors. Indeed, empirical research has consistently shown that specific melanosome dimensions, particularly thin, elongated or hollow melanosomes are associated with iridescent coloration ( 7-9 ). While the physical mechanisms of iridescent color production have been intensively investigated ( e.g., 3,5,6,10-12 ), how iridescence evolved remains less clear and is mostly limited to single clades such as ducks, cuckoos and sunbirds ( e.g.,7,11-15, but see 16 ). Melanosomes that are thin-elongate rods or of specialised shape that are deposited as layers in a keratin feather matrix, were present in avian and non-avian dinosaurs from the Jurassic ( 17 ) and the Cretaceous ( Microraptor , Eoconfuciusornis , Wulong bohaiensis ) ( 18-20 ), suggesting that iridescence was present in the earliest ancestors of birds ( 16 ). However, these studies do not consider the substantial variation in intensity of iridescent coloration that ranges from the subtle sheen of swifts to the brilliant, intense iridescence of peacocks. Nor do they take the topology of coloration, i.e. where iridescence appears on the body, into account, even though this likely influences its function and how it evolves or intensifies ( 13,21-26 ). Close to 70 years ago, pioneering researchers ( 3,27 ) established a classification scheme with four different degrees of iridescence (weak, moderate, strong, brilliant) (See material and methods for definitions). These categories correspond to colors represented by, for example, swifts (weak), many corvids (moderate), ducks (strong) and hummingbirds (brilliant) (Figure 1A). These groups have been previously used and validated ( 6,15,28 ), and new measurements on 105 specimens confirm that they are repeatable and quantifiable (see below). Here we scored the intensity of iridescence (weak to brilliant) in 71,536 body patches from 8,942 specimens of 5,755 bird species (covering 98% of genera and all families ( 29 )) to investigate how and when iridescence, and differences in intensity, evolved in birds and extrapolate these results to non-avian dinosaurs. Results and Discussion Gradual evolution of iridescence from both melanized and non-melanized feather patches. Iridescence was observed in 123 of 193 bird families (64%) (Supplementary data 1,2; Figure 1; Supplementary figure 1-2), substantially expanding the number of families previously (n=85) known to exhibit iridescence ( 3,6,15,27, but see 16 , which only included iridescence stronger than weak, i.e. at least moderate). Furthermore, iridescence is present in 1,793 species (30% of species sampled) and 654 genera (30% of genera sample) (Supplementary data 1). Even though iridescence produces some of the most conspicuous colors, it is much more common than previously realized. To identify transition patterns, we modelled the relative frequencies of gains and losses under five different evolutionary models (Supplementary table 1). These analyses indicated that overall, ordered models perform best (Supplementary table 2-3). More specifically, ordered 2 but also ordered 1 had the highest AIC values across most datasets. While ordered 1 enforces all gains within iridescence classes (i.e. scores > 1) to be gradual (i.e. states cannot jump over adjacent iridescent states), ordered 2 is less stringent with a single constraint: brilliant iridescence only evolves out of strong iridescence. This suggests that transitions between different classes of iridescence—and between gains and losses—do not occur at equal frequencies. More specifically, they suggest that iridescence, in particular the evolution of brilliant iridescence, is highly gradual. Averaged across all models and patches, rates of evolution (Figure 2) from non-adjacent states (on average, r 0-3→5; male and female, strict and extended ) to brilliant iridescence are 0 or close to it. Similarly, on average, skipping from weak iridescence to strong iridescence (on average, r 2→4; male = 0.004, r 2→4; male_extended = 0.008, r 2→4; female =0.002, r 2→4; female_extended =0.004) is on average three to eight times less likely than strong iridescence evolving out of moderate iridescence (on average, r 3→4; male =0.012, r 3→4; female =0.016, r 3→4 female_extended =0.011), but not in the extended male datasets (r 3→4; male_extended =0.001). Going from non-iridescence towards moderate (r 0,1→3; male < 0.002, r 0,1→3; male_extended <0.009, r 0,1→3; female < 0.002, r 0,1→3; female_extended < 0.004) or strong (r 0,1→4; male =0, r 0,1→4; male_extended =0 , r 0,1→4; female = 0, r 0,1→4; female_extended = 0) iridescence while skipping weak iridescence is also more unlikely than strong (on average, r 3→4; male =0.016, r 3→4; female = 0.016, r 3→4 female_extended =0.011) and moderate (on average, r 2→3; male =0.055, r 2→3; male_extended =0.094, r 2→3; female = 0.054, r 2→3; female =0.076,) iridescence evolving out of their adjacent state, except for the extended male dataset (r 3→4; male_extended =0.001). Similar, often even stronger, patterns of gradual evolution were observed across datasets, patches and perturbations. For example, while, on average, in the male extended dataset it seems that strong iridescence is less likely to evolve out of moderate iridescence compared to even weaker intensities, this exception is lost in the perturbated dataset, and seems to be driven by high r 3→4 rates in crown, throat and belly. These results strongly suggest that the evolution of iridescence was gradual, with non-ordered models being the best model in 3 (all ARD models) out of 72 tested configurations, in which the rates are still mostly consistent with an ordered model (Supplementary Figure 4-11). Such gradual evolution might reflect the need for novel melanosome structures to produce increasingly intense colors, as has been observed in Starlings (Sturnidae) ( 7 ), or further modification of barbules. While iridescence is relatively common across the avian phylogeny, it has been lost more often than gained (Figure 2, 3; Supplementary figures 3-11; Supplementary data 3 – see dryad for files) a pattern previously recovered for Icterids ( 30 ) and Cuckoos ( 15 ) but not for starlings ( 7 ). As predicted, iridescence most frequently evolves out of non-organised melanosomes (on average, r 1→2-5; male < 0.002; r 1→2-5; male_extended < 0.002; r 1→2-5; female < 0.007; r 1→2-5; female_extended < 0.006). However, while melanin is essential to produce most iridescent feathers, our ancestral state estimations showed that for multiple patches (and consistently across datasets) transition rates for non-melanized (on average, r 0→2-5; male <0.002; r 0→2-5; male_extended <0.004; r 0→2-5; female <0.002; r 0→2-5; female_extened <0.004) states are not zero, meaning that the presence of (disorganised) melanin in the ancestral patch is not essential for iridescence to evolve. That is, ordered melanin can evolve out of a feather with or without disordered melanin. Indeed, the existence of non-melanosome-based iridescence in the feathers of manakins and tanagers ( 31, 32 ); but also in other integumentary types such as the beak ( 33 ), are consistent with the results observed here. Alternatively, melanosome-based iridescence might also evolve out of non-melanized barbules. While this occurred many times, one clear example is Leptocoma where putative ancestral L. minima , L. brasiliana and L. zeylonica had white or carotenoid-bellies, while the derived L. aspasia and L. calcostetha have iridescent belies. Previous studies did not find such results (i.e. iridescence evolved out of melanized feathers, e.g., 30 ) likely stemming from the exclusion of non-melanized species in earlier analyses. Our results thus suggest that other factors in addition to the presence of melanosomes are important for the evolution of iridescent colors, such as the dimensions and elongation of the barbule ( 10, 15, 30, 34, 35 ). Flattening of the barbule provides more surface area for reflection and, when combined with increasing melanosome density, could promote passive self-assembly and reorganization of melanosomes into ordered nanostructures ( 30 ). Whether barbule morphology contributes in this way to the likelihood of iridescence evolution is a fascinating question for future research. Finally, patches do not evolve in isolation, so iridescence in one area may be influenced by the presence of nearby melanized feathers ( 14 ). However, the maximum iridescent value across all patches (i.e. the highest score attributed to any patches per species), which eliminates the effect of scoring individual patches by reducing each species to a single maximum score, shows similar rates from non-melanized and melanized feathers to iridescent feathers (e.g., r 1→2-5; male =0-0.007; r 1→2-5; male_extened =0-0.006; r 1→2-5; female =0-0.007; r 1→2-5; female_extened =0-0.005 and r 0→2-5; male =0-0.005; r 0→2-5; male_extened =0-0.003; r 0→2-5; female =0.0-0.005; r 0→2-5; female_extened =0-0.003). As such, the observed pattern does not seem to be an artifact associated with patch delimitation and is consistent with the evolution of iridescence from non-melanized plumage. Brilliant iridescence is ancestral to birds In addition to providing evolutionary transition rates, ASEs also provide us with probability estimates for different iridescent intensities in different nodes (i.e. ancestors). These ASE analyses (Figure 3-4, Supplementary figure 12, Supplementary data 3) show that the male most recent common ancestor (MRCA) of birds was, with a probability higher than 95%, brilliantly iridescent (extended male dataset), in multiple patches (Figure 3; Supplementary table 4), or at least strong iridescent (male and female datasets). The tendency of the extended dataset to produce higher iridescence scores is potentially linked to an increase in power by including sexually monomorphic iridescent species. Males are generally more iridescent than females, thus while the inclusion of mis-sexed specimens could potentially inflate iridescence intensities in the female extended dataset, this is not the case in the male extended dataset. This makes the extended dataset, with more specimens and no overestimation, the most reliable estimations, with brilliant iridescent in the wings, throat and breast in the MRCA of birds and varying degrees of iridescence in other patches (Figure 4). These results hold in a perturbated datasets where iridescence is artificially, on average, down-scored. Given that extant paleognaths are non-iridescent, an iridescent MRCA seems counterintuitive, however, fossil paleognaths (Lithornithidae), were likely iridescent, consistent with the findings here ( 36 ). Previous work ( 20 ) suggested that iridescence was present in the MRCA of birds, but did not quantify in which patches iridescence occurred, nor the intensity of iridescence. Such brilliant iridescent colors can today be observed in spectacular birds such as hummingbirds and birds-of-paradise and often require specialized melanosomes. Furthermore, as iridescence evolves gradually, brilliant iridescence requires an ancestral species with at least strong iridescence, suggesting that some dinosaurs, had at least strong iridescence, and potentially brilliant iridescence, in some body patches. Most dinosaur color reconstructions rely on the preservation of melanosomes and the known relationship between melanosome shape and coloration ( 17, 19, 37-41 ). As such, these studies show that dinosaur (and pterosaur) feathers were black, brown and reddish-brown. However, these findings fail to provide a broader picture of color evolution and only reveal a limited subset of color space compared to that present in extant birds ( 42 ). Given the broad range of colors in birds and virtually all other tetrapod groups, it is unlikely that dinosaurs had only dull melanin-based colors such as brown and black. Thin and flattened melanosomes with clear signs of layering, previously discovered in dinosaurs from the Jurassic ( 17 ) and Cretaceous ( 18 ), suggest the presence of at least weak but potentially more intense iridescence in dinosaurs ( 19 ). As such, the results of our ASE align with previously identified iridescence in dinosaurs, including differential distribution of iridescence, e.g. as in Wulong bohaiensis (occurring 30 mya after the split between avian and non-avian dinosaurs, suggesting the presence of this pattern well before this split) ( 19 ), but our results add a quantification of its intensity, with brilliant iridescent colors likely present in the ancestor of birds, and potentially dinosaurs (Figure 4). While these brilliant iridescent colors expand the potential color space in plumages from just that of white, red, and black it is possible that non-avian dinosaur color space was even larger. Putative fossilized carotenoids have not been described and ASE using extant birds strongly suggest that an absence of fossilized carotenoids might correspond to a general absence of carotenoids in dinosaur feathers ( 43 ). As such, structural color might indeed have been a good alternative for generating patterns and color for the purpose of signaling. Nevertheless, other rare, clade-specific pigments (penguins, turacins, turacoverdins and psittacofulvins) are found in birds ( 44-46 ), and it is plausible that extinct “ghost pigments”, i.e. unknown pigments that left no trace in the fossil record, were present in dinosaurs further expanding color space. Different iridescent classes are perceivable The usage and validity of quantitative, rather than qualitative, classification of coloration has been discussed extensively before ( 47-49 ). Indeed, a potential caveat of this study is the scoring of iridescent classes itself. Yet, colors can be consistently assigned to distinct categories, even across cultures ( 50-51 ), and qualitative studies have provided valuable insights when quantification is difficult ( 20, 52, 53 ). By revisiting and re-scoring 123 species previously scored by Durrer ( 3 ) and Auber ( 27 ), we show that there is a strong correspondence between classifications decades apart, even when accounting for obvious misidentifications (see methods (Supplementary data 4). The similarity between classifications ranges from 78 to 85% depending on the inclusion/exclusion of such misidentifications (Supplementary data 4; Supplementary figure 13). Mismatches almost exclusively occurred between subsequent categories (e.g., weak vs. moderate; strong vs. brilliant). Furthermore, we used reflectance spectrophotometry (n=105) (Supplementary data 5) to objectively measure color and identify the mechanisms behind these different categories. These analyses that chroma and contrast, two key-features of iridescent reflectance curves–differ significantly among all categories (Supplementary table 5; Supplementary figure 14) (p-values < 0.007 between all categories). This is expected, given that iridescence produces a distinct, often multi-peak reflectance curve, unlike other color mechanisms. Indeed, chroma and contrast both incorporate the difference between maximum and minimum reflectance, which is expected to be higher in more “peaky” reflectance spectra where lows correspond to the high absorption by melanin and highs correspond to the selective reflectance. Hue (the observed color) and brightness (overall reflectance), however, are significantly different between some, but not all categories. The absence of a strong signal might indicate that all mechanisms can have overall similar brightness levels (e.g. carotenoids can also be bright), and that some mechanisms (e.g. brilliant iridescence) can produce colors that overlap with both carotenoids and other iridescence intensities. Additional support comes from ordered logistic regressions where we show that precision, recall and F1-score (i.e. the harmonic mean of precision and recall) in a model based on reflectance data where between 2 and 6 times higher than a random model (Supplementary table 6 and 7). Non-perfect scores are potentially attributed to measurement error, or color-metrics visible (in 3D, e.g. by moving the specimen) but not measured using point measures. Indeed, we measured iridescence at one specific angle because we classified iridescence based on the most intense color. Potentially this is also influenced by relative change in color between angles. Nonetheless, classification performance increased from weak to brilliant iridescence, and misclassifications were mostly underestimation, suggesting that our results are conservative. Finally, using a perturbated dataset, where mis-scorings were introduced based on the empirically observed error-rates (Supplementary data 4), we re-performed all analyses and found no difference across analyses (see discussion throughout text). These results indicate that human eyes can distinguish between different classes of iridescence intensity. Notably, iridescence intensity, largely independent of hue (Supplementary table 5, Supplementary figure 14), is a continuous trait but can be categorized into discrete classes. Indeed, previous research used similar terms (weak, strong, brilliant) ( 6, 15, 28 ) and found quantifiable support for these groups ( 15 ). Such categorization suggests that hue and intensity may represent distinct evolutionary and developmental modules that can evolve independently, potentially optimizing traits such as visibility. While shifting hue to align with spectral sensitivities may enhance species-specific conspicuousness, reducing iridescence intensity could, in turn, lower overall conspicuousness. However, this effect seems to apply primarily to weak-to-strong iridescence transitions, as highly brilliant colors at shorter wavelengths appear to be rare—though exceptions exist, such as in peacocks and the bird of paradise ( Lophorina ), which were not included in our measurements but show hues consistent with low wavelengths. Conclusions Birds display some of the most elaborate ornaments in nature, often relying on melanosome-based iridescent colors for visual communication. This study shows that iridescence can evolve directly from non-melanized feathers, without passing through a stage of non-iridescent melanin-based coloration, suggesting that melanin alone is not sufficient and other barbule traits play a crucial role. Even a disorganized layer of melanosomes within a keratin cortex can produce weak iridescence, hinting at a low barrier to the origin of this trait. Finally, the ancestral presence of iridescence in bird feathers, imply that iridescence may have been widespread in extinct species like dinosaurs and pterosaurs, potentially making the Mesozoic era far more colorful than previously thought. Material and Methods Specimen collection and iridescence assessment We scored the presence of iridescence in 71,536 body patches, from 8,942 representative specimens (round skins) of 5,755 species of birds (more than 50% of all species) encompassing all but 36 genera (98% complete) and all families (29). We scanned all specimens available and chose a representative male (n=4263) and female (n=3618) of each species. When sex was unknown, we scored the unsexed specimen (n=1060). When multiple morphs were present, we used the most iridescent morph available. We investigated iridescence intensity (i.e. how conspicuous an iridescent color is), not iridescence per se (i.e. how much do colors change with changing angles). For each species we scored the presence and intensity of iridescence in eight patches corresponding to known feather tracts: crown, mantle, rump, tail, wings, throat, breast and belly. Non-iridescent colors were scored as either melanin-based colors (grey and black colors, as well as non-carotenoid brown colors), or as non-melanin-based colors. Iridescence was further scored by intensity, following Auber (1953) (27) and Durrer (1977) (3), where we scored iridescence as weak when “only close inspection reveals iridescence”, moderate when “the primary impression given by the feathers in situ is black – Rook Corvus frugileges”), strong (“head of Mallard Anas platyrhyncos”) and brilliant (“many Trochilidae”). We created four datasets; All hypotheses were tested for the two sexes (female dataset n=3610; male dataset n=4256); separately, as selective forces can act independently, and even divergently, in different sexes (25, 29-34). Additionally, to analyze datasets containing all species, we conducted analyses using the “extended” dataset, which comprises both male (male_extended and female_extended, n=5742) and female specimens. These datasets were based on the original male and female datasets but included additional female, male and/or unsexed specimens such that each species sampled was present in the extended dataset, even if the target sex was absent. To verify this apparent subjective classification, we compared our scores with those of Auber and Durrer for 123 species and calculated overlap in score assignment between scorers. Next, we compared our scores with those of Auber and Durrer for 123 species (supplementary figure 13) and found that iridescent scores were highly repeatable (Supplementary discussion, Supplementary figure 13). To further confirm the validity of the Auber and Durrer classification, we compared reflectance spectra of 236 specimens from 105 species with known Auber and Durrer classifications (Supplementary data 3). These were measured at the Museum of Comparative Zoology (supplementary data 4) using an Ocean Optics USB2000 spectrophotometer measuring UV-VIS (300-700nm) (calibrated with a dark and white standard) with an PX2 Light Source. We connected the fibre optic cable with a reflection probe holder that was held at an angle of 90º placed directly on the feather patch, with the probe positioned at 0.5 cm from the sample. As we were measuring iridescent patches, we measured at the angle of maximal reflection. We only measured iridescent patches to compare them to the Auber and Durrer classification. Given that older specimens (>50 years) (54) might have altered colors, we took precautionary measures (55) and excluded specimens that showed physical damage and dusting. Three measurements per patch were taken, negative values were removed and spectra were smoothed using the R package ‘pavo’ (56). To quantitatively and objectively measure iridescent colors of the different classes proposed by Auber and Durrer, we calculated summary statistics per species and used phylogenetic ANOVA implemented using phytools’s (57) phylANOVA to compare differences in brightness and chroma and saturation between categories (supplementary figure 14, supplementary table 5). We tested multiple variables using pavo: B2 (Mean brightness; Mean relative reflectance over the entire spectral), B3 (Intensity: Maximum relative reflectance), S3 (Chroma: Reflectance over the Rmax +- 50nm range divided by B1, i.e. the sum of the relative reflectance over the entire spectral range), S6 (Contrast: Rmax – Rmin), S10 (Peaky chroma: (Rmax - Rmin)/B2), H2 (Hue: Wavelength at bmaxneg), and finally S9 (Carotenoid chroma: (R700 - R450)/R700) as a control. These showed that there were consistent, and detectable differences across iridescence categories (weak, moderate, strong, brilliant), allowing the subsequent use in the downstream analyses. In addition, to calculate precision, recall and F1 (i.e. the harmonic mean of precision and recall) we performed an ordered logistic regression implemented in MASS (58) with the reflectance data summary parameters as predictors. These results (Supplementary table 6-7) showed that our model outperforms a random model. Nonetheless, we accounted for uncertainty in the analyses below by including perturbated datasets, in which we randomly changed iridescent scores based on the misclassifications between Durrer, Auber, and this study. By using these empirically observed misclassification rates (Supplementary Figure 13), we exclude unlikely misidentifications (e.g. weak for brilliant), but still introduce up to 50% variability (e.g. in weak iridescence where most misidentifications occur). Ancestral state estimations (ASE) All analyses (here and in the sections below) were run separately for males and females, as well as a male and female “extended” dataset, that included all male/female specimens but were expanded with species for which we either had no information on the sex, or specimens were only available for one sex. All analyses were run on the most recent all-bird phylogeny (59). For ASE specifically, we estimated the evolution of all eight patches, as well as the maximum degree of iridescence observed across all patches using fitMk in phytools (57). We used fitMk to test five different evolutionary models, the symmetric rates (SYM), the equal rates (ER), the all-rates-different model (ARD), and two ordered models. More specifically, ordered 1 is an ARD model that enforces all gains within iridescence classes (i.e. scores > 1) to be gradual (i.e. states cannot jump over adjacent iridescent states), ordered 2 is less stringent with a single constraint: brilliant iridescence only evolves out of strong iridescence. We compared models using AIC values. Ancestral states were then estimated using the ancr function, which computes marginal ancestral states. All analyses were rerun on perturbated datasets. Declarations Acknowledgments: We thank the EON lab for discussing the results. We also thank Annelore Nackaerts from RMCA, Olivier Pauwels from RBINS, Jeremiah Trimble and Kate Eldridge from MCZ for access to their collections. Finally, this work would have been impossible without the multiple generations of scientists that have collected the birds that were used to quantify iridescence. Funding: B.A.E.F. (Belgian American Educational Foundation) (M.P.J.N) FWO (Fonds Wetenschappelijk Onderzoek) (M.P.J.N) UGent BOF mandate (M.P.J.N) European Research Council under the European Union’s Horizon 2020 research and innovation program; grant agreement No. 101000504 (M.C.) Portuguese Foundation for Science and Technology (FCT, https://www.fct.pt) research contract CEECINST/00014/2018/CP1512/CT0002 (M.C.). Author contributions: Conceptualization: M.P.J.N., R.C.K.B., S.V.E., M.C., L.D.A., M.D.S. Methodology: M.P.J.N., L.D.A., M.D.S. Investigation: M.P.J.N. with help from G.D. Visualization: M.P.J.N. Supervision: R.C.K.B., S.V.E., M.C., L.D.A., M.D.S. Writing – original draft: M.P.J.N. Writing – review & editing: all authors. Competing interests: There are no competing interests. Data and materials availability: All data and code to reproduce the results are available in the main text or the supplementary materials. References G.E., Hill, G. E., K. J., McGraw, (Eds.), Bird coloration (Vols. 1–2). (2006, Harvard University Press). L., D’Alba, M.D., Shawkey, M. D. Melanosomes: Biogenesis, properties, and evolution of an ancient organelle , Physiological Reviews 99 (1), 1-19, (2019). H., Durrer. Schillerfarben der Vogelfeder als Evolutionsproblem. (1977, University of Basel). S. M. Doucet, M. D. Shawkey, G. Hill, R. Montgomerie, Iridescent plumage in satin bowerbirds: structure, mechanisms and nanostructural predictors of individual variation in color, Journal of Experimental Biology 209 , 380-390 (2006). P. Freyer, B. D. Wilts, D. G. Stavenga, Cortex thickness is key for the colors of iridescent starling feather barbules with a single, organized melanosome layer, Frontiers in Ecology and Evolution 9 , 746254 (2021). K. K. Nordén, C. M. Eliason, M. C. Stoddard, Evolution of brilliant iridescent feather nanostructures, eLife 10 , e71179 (2021). R. Maia, D. R. Rubenstein, M. D. Shawkey, Key ornamental innovations facilitate diversification in an avian radiation, Proceedings of the National Academy of Sciences U.S.A. 110 , 10687-10692 (2013). H. Durrer, W. Villiger, Schillerfarben der Trogoniden, J. Ornithol. 107, 1-26 (1966). D. G. Stavenga, H. L. Leertouwer, B. D. Wilts, Magnificent magpie colors by feathers with layers of hollow melanosomes, Journal of Experimental Biology 221 , jeb174656 (2018). R. Maia, R. H. Macedo, M. D. Shawkey, Self-assembly of iridescent barbules through depletion attraction of melanosomes during keratinization, Journal of the Royal Society Interface 9 , 734-743 (2011). C. M. Eliason, R. Maia, J. L. Parra, M. D. Shawkey, Signal evolution and morphological complexity in hummingbirds (Aves: Trochilidae), Evolution 74 , 447-458 (2020). C. M. Eliason, R. Maia, M. D. Shawkey, Modular color evolution in birds facilitated by a complex nanostructure, Evolution 69 , 357-367 (2015). M. P. J. Nicolaï, S. Rogalla, M. Yousefi, R. C. K. Bowie, L. D'Alba, M. D. Shawkey, Ecological, genetic and geographical divergence explain differences in coloration among sunbird species (Nectariniidae), Ecology and Evolution 14 , e11427 (2024). M. P. J. Nicolaï, B. Van Hecke, S. Rogalla, G. Debruyn, R. Bowie, N. J. Matzke, S. Hackett, L. D'Alba, M. Shawkey, The evolution of multiple color mechanisms is correlated with diversification in sunbirds (Nectariniidae), Systematic Biology syae006 (2024). K. K. Nordén, C. R. Cooney, F. Babarovic, M. C. Stoddard, Historical contingency influences the diversity of feather nanostructures in cuckoos, bioRxiv (2023). M. Eliason, M. P. J. Nicolaï, C. Bom, E. Blom, L. D'Alba, M. D. Shawkey, Transitions between color mechanisms affect speciation dynamics and range distributions of birds, Nature Ecology and Evolution 8 , 1723-1734 (2024). D. Hu, J. A. Clarke, C. M. Eliason, R. Qiu, Q. Li, M. D. Shawkey, C. Zhao, L. D'Alba, J. Jiang, X. Xing, A bony-crested Jurassic dinosaur with iridescent plumage highlights complexity in early paravian evolution, Nature Communications 9 , 217 (2017). Q. Li, K. Gao, Q. Meng, J. A. Clarke, M. D. Shawkey, L. D'Alba, R. Pei, M. Ellison, M. A. Norell, J. Vinther, A new specimen of Microraptor and the evolution of iridescent plumage color, Science 335 , 1215-1219 (2012). Y. Pan, Z. Li, M. Wang, T. Zhao, X. Wang, X. Zheng, Unambiguous evidence of brilliant iridescent feather color from hollow melanosomes in an Early Cretaceous bird, National Science Reviews 9 , nwab227 (2021). D. Croudace, C. Shen, J. Lü, S. M. Brusatte, J. Vinther, Iridescent plumage in a juvenile dromaeosaurid theropod dinosaur, Acta Palaeontologica Polonica 68 , 213-225 (2023). S. M. Doucet, M. G. Meadows, Iridescence: A functional perspective, Journal of the Royal Society Interface 6 , S115-S132 (2009). K. Kjernsmo, H. M. Whitney, N. E. Scott-Samuel, J. R. Hall, H. Knowles, L. Talas, I. C. Cuthill, Iridescence as camouflage, Current Biology 30 , 551-555 (2019). R. K. Simpson, A. F. Mistakidis, S. M. Doucet, Natural and sexual selection shape the evolution of color and conspicuousness in North American wood-warblers (Parulidae), Biological Journal of the Linnean Society 130 , 89-100 (2020). D. Gomez, M. Théry, Simultaneous crypsis and conspicuousness in color patterns: Comparative analysis of a Neotropical rainforest bird community, American Naturalist 169 , S42-S61 (2007). R. S. Marcondes, R. T. Brumfield, Fifty shades of brown: Macroevolution of plumage brightness in the Furnariida, a large clade of drab neotropical passerines, Evolution 73, 704-719 (2019). F. Babarović, C. R. Cooney, Z. K. Varley, L. O. Nouri, N. J. Nadeau, G. H. Thomas, The effects of ecology and behaviour on the evolution of coloration in Coraciiformes, Behavioral Ecology 34 , 387-398 (2023). L. Auber, The distribution of structural colors and unusual pigments in the class Aves, Ibis 99 , 463-476 (1957). D. Stuart-Fox, L. Ospina-Rozo, L. Ng, A. M. Franklin, The paradox of iridescent signals, Trends in Ecology and Evolution 36 , 187-195 (2021). J. F. Clements et al., The eBird/Clements checklist of birds of the world: v2018 (2018). M. D. Shawkey, M. E. Hauber, L. K. Estep, G. E. Hill, Evolutionary transitions and structural mechanisms of avian plumage coloration in grackles and allies (Icteridae), Journal of the Royal Society Interface 3 , 777-783 (2006). B. Igic, L. D'Alba, M. D. Shawkey, Manakins produce bright iridescent feather colors without melanosomes, Journal of Experimental Biology 219 , 1851-1859 (2016). D. C. Skigin, M. E. Inchaussandague, C. D'Ambrosio, A. Barreira, P. Tubaro, How the observed color of the Swallow Tanager (Tersina viridis) changes with viewing geometry, Optik 182 , 639-646 (2019). M. Nicolaï, G. Debruyn, M. Soenens, M. D. Shawkey, L. D'Alba, Nanoscale millefeuilles produce iridescent bill ornaments in birds, PNAS Nexus pgae138 (2024). J.-P. Iskandar, C. M. Eliason, T. Astrop, B. Igic, R. Maia, M. D. Shawkey, Morphological basis of glossy red plumage colors, Biological Journal of the Linnean Society 118 , 106-118 (2016). L. D'Alba, M. Meadows, M. Maia, J. S. Yeo, M. Manceau, M. D. Shawkey, Morphogenesis of iridescent feathers in Anna's hummingbird Calypte anna, Integrative and Comparative Biology icab123 (2021). E.M. Eliason, J.A. Clarke, Cassowary gloss and a novel form of structural color in birds. Sci. Adv. 6,eaba0187(2020). Q. Li, G. Keqin, J. Vinther, M. D. Shawkey, J. A. Clarke, L. D'Alba, Q. Meng, D. E. G. Briggs, R. O. Prum, Plumage color patterns of an extinct dinosaur, Science 327 , 1369-1372 (2010). R. Carney, J. Vinther, M. D. Shawkey, L. D'Alba, J. Ackermann, New evidence on the color, ultrastructure, and nature of the isolated Archaeopteryx feather, Nature Communications 3 , 637 (2012). J. Vinther, D. E. G. Briggs, R. O. Prum, V. Saranathan, The color of fossil feathers, Biology Letters 4 , 522-525 (2008). K. K. Nordén, J. W. Faber, F. Babarović, T. L. Stubbs, T. Selly, J. D. Schiffbauer, P. Peharec Štefanić, G. Mayr, F. M. Smithwick, J. Vinther, Melanosome diversity and convergence in the evolution of iridescent avian feathers-Implications for paleocolor reconstruction, Evolution 73 , 15-27 (2019). Cincotta, M. Nicolaï, H. B. N. Campos et al., Pterosaur melanosomes support signalling functions for early feathers, Nature 604 , 684-688 (2022). M. C. Stoddard, R. O. Prum, How colorful are birds? Evolution of the avian plumage color gamut, Behavioral Ecology 22 , 1042-1052 (2011). S. N. Davis, J. A. Clarke, Estimating the distribution of carotenoid coloration in skin and integumentary structures of birds and extinct dinosaurs, Evolution 76 , 42-57 (2022). D. B. Thomas, C. M. McGoverin, K. J. McGraw, H. F. James, O. Madden, Vibrational spectroscopic analyses of unique yellow feather pigments (spheniscins) in penguins, Journal of the Royal Society Interface 10 , 20121065 (2013). J. Van Dyck, Reflectance spectra of plumage areas colored by green feather pigments, Auk 109 , 293-301 (1992). R. Arbore, S. Barbosa, J. Brejcha, Y. Ogawa, Y. Liu, M. P. J. Nicolaï, P. Pereira, S. J. Sabatino, A. Cloutier, E. S. K. Poon, C. I. Marques, P. Andrade, G. Debruyn, S. Afonso, R. Afonso, S. G. Roy, U. Abdu, R. J. Lopes, P. Mojzeš, M. Carneiro, A molecular mechanism for bright color variation in parrots, Science 386 , eadp7710 (2024). J. A. Endler, On the measurement and classification of color in studies of animal color patterns, Biological Journal of the Linnean Society 41 , 315-352 (1990). Y. He, Z. K. Varley, L. O. Nouri et al., Deep learning image segmentation reveals patterns of UV reflectance evolution in passerine birds, Nature Communications 13 , 5068 (2022). K. Delhey, M. Valcu, C. Muck, J. Dale, B. Kempenaers, Evolutionary predictors of the specific colors of birds, Proc. Natl. Acad. Sci. U.S.A. 120, e2217692120 (2023). E. Gibson, R. Futrell, J. Jara-Ettinger, K. Mahowald, L. Bergen, S. Ratnasingam, M. Gibson, S. T. Piantadosi, B. R. Conway, Color naming across languages reflects color use, Proceedings of the National Academy of Sciences U.S.A. 114 , 10785-10790 (2017). N. Seddon, J. A. Tobias, M. Eaton, A. Ödeen, Human vision can provide a valid proxy for avian perception of sexual dichromatism, Auk 127, 283-292 (2010). M. P. J. Nicolaï, M. D. Shawkey, S. Porchetta, L. D'Alba, Exposure to UV radiance predicts repeated evolution of concealed black skin in birds, Nature Communications 11 , 2414 (2020). M. P. J. Nicolaï, R. Vanisterbecq, M. D. Shawkey, L. D'Alba, Back in black: Melanin-rich skin color associated with increased net diversification rates in birds, Biology Letters 19 , 20230304 (2023). J. K. Armenta, P. O. Dunn, L. A. Whittingham, Effects of specimen age on plumage color, Auk 125 , 803-808 (2008). S. M. Doucet, G. E. Hill, Do museum specimens accurately represent wild birds? A case study of carotenoid, melanin, and structural colors in long–tailed manakins Chiroxiphia linearis, Journal of Avian Biology 40 , 146-156 (2009). R. Maia, H. Gruson, J. A. Endler, T. E. White, pavo 2: New tools for the spectral and spatial analysis of color in R, Methods in Ecology and Evolution 10 , 1097-1107 (2019). L. J. Revell, phytools 2.0: An updated R ecosystem for phylogenetic comparative methods (and other things), PeerJ 12 , e16505 (2024). W. N. Venables, B. D. Ripley, Modern Applied Statistics with S , 4th ed. (Springer, New York, 2002). E. J. McTavish, J. A. Gerbracht, M. T. Holder, M. J. Iliff, D. Lepage, P. Rasmussen, B. Redelings, L. L. Sanchez Reyes, E. T. Miller, A complete and dynamic tree of birds, bioRxiv 2024.05.20.595017 (2024). Additional Declarations There is NO Competing Interest. Supplementary Files Supplementarytable1Transititionratemodelstested.xlsx Supplementary table 1 - Transitition rate models tested Supplementarytable2Modelancestralstateestimationssupportvaluesfororiginaldatasets.xlsx Supplementary table 2 - Model (ancestral state estimations) support values for original datasets Supplementarytable3Modelancestralstateestimationssupportvaluesfororiginaldatasetspert.xlsx Supplementary table 3 - Model (ancestral state estimations) support values for perturbated datasets Supplementarytable4Ancestralstateestimationsforallmodels.xlsx Supplementary table 4- Ancestral state estimations for all models Supplementarydata1Iridescenttaxa.csv Supplementary data 1 Supplementarydata2finaliridescencev2withmetadata.csv Supplementary data 2 supp3.zip Supplementary data 3 Supplementarydata4.xlsx Supplementary data 4 Supplementarydata5.txt Supplementary data 5 Supplementaryresults.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8020267","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":539726841,"identity":"eae022ec-680c-4637-b66b-e40abd5842bf","order_by":0,"name":"Michael Nicolai","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAklEQVRIie3PsWrDMBCA4QODujjpeiHBeYLAGYOTQh9Gnjx19xCSQMFdQvfOfYE8gkHgyaDVQwaHgqYMmUKXhlqNTekgm26l6B/EIfQhCcBm+4s5wOp1oQeACtBr9rM+glfCAYMGdBBoCXwRiDZ9ZPbkqOo9QZjfOIeKLxfxTr7mbyfYeyYSCjb3twXC3SMLiOf4sCtVTBmowExchoMUgfTAmSZFiBmI6wsNZPRxackFYyrlWZN1FxkPNg2JUuQkt0wT3vGXcDzJ0SVR/yV6Rv+lZCEWpHzjLVKo0XF575EUh+p0Xk2H9Q4myX5quqXN/R5RP4n6wI9us18dt9lstv/fJ8O8VSkfyQtJAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-9570-0311","institution":"Ghent University","correspondingAuthor":true,"prefix":"","firstName":"Michael","middleName":"","lastName":"Nicolai","suffix":""},{"id":539726842,"identity":"98e102f9-cb45-457b-9b7a-54efb7443e40","order_by":1,"name":"Gerben Debruyn","email":"","orcid":"","institution":"Ugent","correspondingAuthor":false,"prefix":"","firstName":"Gerben","middleName":"","lastName":"Debruyn","suffix":""},{"id":539726843,"identity":"e6707f33-6789-4654-8d7b-785ac5ce263c","order_by":2,"name":"Rauri Bowie","email":"","orcid":"https://orcid.org/0000-0001-8328-6021","institution":"University of California, Berkeley","correspondingAuthor":false,"prefix":"","firstName":"Rauri","middleName":"","lastName":"Bowie","suffix":""},{"id":539726844,"identity":"dc545602-4e28-4c16-861f-0149da739e6f","order_by":3,"name":"Scott Edwards","email":"","orcid":"https://orcid.org/0000-0003-2535-6217","institution":"Harvard University","correspondingAuthor":false,"prefix":"","firstName":"Scott","middleName":"","lastName":"Edwards","suffix":""},{"id":539726845,"identity":"b7292e97-bfb3-4d00-9581-8b9e92e56acf","order_by":4,"name":"Miguel Carneiro","email":"","orcid":"https://orcid.org/0000-0001-9882-7775","institution":"BIOPOLIS - University of Porto","correspondingAuthor":false,"prefix":"","firstName":"Miguel","middleName":"","lastName":"Carneiro","suffix":""},{"id":539726846,"identity":"a08664ae-73de-44e1-8de8-125a229bf612","order_by":5,"name":"Liliana D'Alba","email":"","orcid":"","institution":"Naturalis","correspondingAuthor":false,"prefix":"","firstName":"Liliana","middleName":"","lastName":"D'Alba","suffix":""},{"id":539726847,"identity":"8825688d-f182-4aae-89e2-34bab3fe0160","order_by":6,"name":"Matthew Shawkey","email":"","orcid":"https://orcid.org/0000-0002-5131-8209","institution":"University of Ghent","correspondingAuthor":false,"prefix":"","firstName":"Matthew","middleName":"","lastName":"Shawkey","suffix":""}],"badges":[],"createdAt":"2025-11-03 14:36:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8020267/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8020267/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":95276387,"identity":"d7a7d019-b7b4-470a-9479-1a0b464adc84","added_by":"auto","created_at":"2025-11-06 08:24:02","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":4269463,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe distribution of the degree of iridescence across body patches in birds.\u003c/strong\u003e \u003cstrong\u003e(A\u003c/strong\u003e) Examples of the different degrees of iridescence; Weak (Chiroxiphia pareola), moderate (Corvus corone), strong (Anas platyrhynchos) and brilliant (Calypte costae). Non-melanized (0) and melanized but not iridescent (1) are not shown. The phylogeny for males \u003cstrong\u003e(B)\u003c/strong\u003e (n=5754) and females \u003cstrong\u003e(C)\u003c/strong\u003e (n=5754) is surrounded by eight circles, each representing the presence or absence of different degrees of iridescence across male body patches (extended dataset; for other datasets, see supp. figure 1-2). The degrees of iridescence—weak, moderate, strong, and brilliant—are depicted in varying shades of purple (states 2-5), while grey (state 1) and orange (state 2) represent melanin-based and non-melanized colors, respectively. Node labels on the phylogeny correspond to 14 bird clades, for which ancestral states are available in figure 3; 1) MRCA of birds, 2) Paleognathae, 3) Galloanseres, 4) Mirandornithes, 5) Strisores, 6) Aequornithes, 7) Neoaves, 8) Afroaves, 9) Australaves, 10) Passeriformes, 11) Eurylaimides, 12) Tyrannides, 13) Passeri, 14) Passerides, 15) Corvides.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/b0476c973fdea07687f57b96.png"},{"id":95276398,"identity":"1fea8d52-4213-4bc0-abcb-f76d50d1566b","added_by":"auto","created_at":"2025-11-06 08:24:05","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1658051,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEvolutionary transitions and distribution of the degree of iridescence across body patches in birds.\u003c/strong\u003e Transition rate estimates between iridescence states (based on ARD model) for \u003cu\u003emales\u003c/u\u003e(left) and \u003cu\u003efemales\u003c/u\u003e (right) in the \u003cu\u003estrict\u003c/u\u003e (top) and \u003cu\u003eextended\u003c/u\u003e(bottom) datasets. Transition rates are averaged across all patches within a dataset. Individual transition rates for individual patches and other datasets, are detailed in supplementary figure 3-11). The degrees of iridescence—weak, moderate, strong, and brilliant—are depicted in varying shades of purple (states 2-5), while grey (state 1) and orange (state 2) represent melanin-based and non-melanized colors, respectively. Perturbation of these datasets (supplementary figure 3) did not change the overall results.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/50e6315329a13446bae3a084.png"},{"id":95276388,"identity":"4230f4e9-c826-4583-96a5-ee9640790288","added_by":"auto","created_at":"2025-11-06 08:24:02","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":2915970,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eAncestral state estimation for degree of iridescence. \u003c/strong\u003e\u003c/em\u003e\u003cem\u003eEstimations are based on the strict (left) and extended (right) dataset showing for each row the respective patch is highlighted for males (top) and females (bottom).\u003c/em\u003e \u003cem\u003eThe degrees of iridescence—weak, moderate, strong, and brilliant—are depicted in varying shades of purple, while grey and orange represent melanin-based and non-melanized colors, respectively\u003c/em\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/d635db50337bdac6b2dbc070.png"},{"id":95276392,"identity":"df3f51f1-1831-4cfb-ba65-70f1a23996d9","added_by":"auto","created_at":"2025-11-06 08:24:04","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":5393516,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eThe evolution of iridescence:\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e Reconstruction of Sinosauropteryx prima (non-iridescent melanin-based colors), Microraptor gui (iridescent colors), and Asteriornis maastrichtensis as a representative early Aves showing iridescent colors of different intensities in different patches. Pavo muticus represent an example of an extant bird with brilliant iridescence. Figure by Mark Witton.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/fb2e93b09e5a51ef661d66b8.png"},{"id":95818678,"identity":"31f3b021-2ff1-4183-8049-64252dc9310d","added_by":"auto","created_at":"2025-11-13 10:28:22","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":14342580,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/a35e81ae-72b2-4d36-83b3-568bdbe55684.pdf"},{"id":95276385,"identity":"2d7a6174-454b-4f1b-9f29-50ae92685d0f","added_by":"auto","created_at":"2025-11-06 08:24:01","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":10464,"visible":true,"origin":"","legend":"Supplementary table 1 - Transitition rate models tested","description":"","filename":"Supplementarytable1Transititionratemodelstested.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/50d1363b8f91833c73581951.xlsx"},{"id":95276384,"identity":"fabfe70d-c8d6-4328-89de-e810c2e92926","added_by":"auto","created_at":"2025-11-06 08:24:01","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":20149,"visible":true,"origin":"","legend":"Supplementary table 2 - Model (ancestral state estimations) support values for original datasets","description":"","filename":"Supplementarytable2Modelancestralstateestimationssupportvaluesfororiginaldatasets.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/5eadfab046795b8703966cf4.xlsx"},{"id":95276390,"identity":"4040eab1-606b-470b-82df-95a807fdde53","added_by":"auto","created_at":"2025-11-06 08:24:02","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":20134,"visible":true,"origin":"","legend":"Supplementary table 3 - Model (ancestral state estimations) support values for perturbated datasets","description":"","filename":"Supplementarytable3Modelancestralstateestimationssupportvaluesfororiginaldatasetspert.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/6acf97462629189d9c9a818b.xlsx"},{"id":95276397,"identity":"eb9cfb45-3294-44f3-8858-a9c0450083f7","added_by":"auto","created_at":"2025-11-06 08:24:05","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":17828,"visible":true,"origin":"","legend":"Supplementary table 4- Ancestral state estimations for all models","description":"","filename":"Supplementarytable4Ancestralstateestimationsforallmodels.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/f7d49fb6704a5e1a2b6d1483.xlsx"},{"id":95276389,"identity":"502475eb-cf8c-4e23-b751-a090ce82b22a","added_by":"auto","created_at":"2025-11-06 08:24:02","extension":"csv","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":37214,"visible":true,"origin":"","legend":"Supplementary data 1","description":"","filename":"Supplementarydata1Iridescenttaxa.csv","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/ed8263586fc380e187d40938.csv"},{"id":95276399,"identity":"8ad66342-6eb4-46c4-aec1-3abc9cff0018","added_by":"auto","created_at":"2025-11-06 08:24:05","extension":"csv","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":1203049,"visible":true,"origin":"","legend":"Supplementary data 2","description":"","filename":"Supplementarydata2finaliridescencev2withmetadata.csv","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/7065401d89a92b5b8f2c5aab.csv"},{"id":95276404,"identity":"7e1a8d27-948f-4812-b1cc-906cefcf4943","added_by":"auto","created_at":"2025-11-06 08:24:10","extension":"zip","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":334092368,"visible":true,"origin":"","legend":"Supplementary data 3","description":"","filename":"supp3.zip","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/2e0429760069e6e0a8901ab1.zip"},{"id":95276396,"identity":"0c041b86-793c-4f8b-a7ea-3ce156a48dcd","added_by":"auto","created_at":"2025-11-06 08:24:05","extension":"xlsx","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":21405,"visible":true,"origin":"","legend":"Supplementary data 4","description":"","filename":"Supplementarydata4.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/ecf21153c35998fcea01c07b.xlsx"},{"id":95276400,"identity":"44f43acf-ac2f-4080-8e9b-ce8f6a195b1c","added_by":"auto","created_at":"2025-11-06 08:24:05","extension":"txt","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":682607,"visible":true,"origin":"","legend":"Supplementary data 5","description":"","filename":"Supplementarydata5.txt","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/3119ad4d72130ee44c145b6f.txt"},{"id":95276393,"identity":"37ad1025-9422-4bdd-bc49-e4253eb1ec43","added_by":"auto","created_at":"2025-11-06 08:24:04","extension":"docx","order_by":10,"title":"","display":"","copyAsset":false,"role":"supplement","size":5791637,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementaryresults.docx","url":"https://assets-eu.researchsquare.com/files/rs-8020267/v1/bbbc2521c8078ca42c02997f.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"\u003cp\u003eThe evolution of brilliant iridescence in birds\u003c/p\u003e","fulltext":[{"header":"Main text","content":"\u003cp\u003eBirds are among the most colorful animals, partly due to the diversity of their color production mechanisms. For example, bird colors can arise from the deposition of various pigments, such as carotenoids, melanins, and protoporphyrins (\u003cem\u003e1\u003c/em\u003e). Additionally, they exhibit iridescent structural colors, which result from light scattering off highly organized nano-scale arrays of melanin-containing organelles called melanosomes (\u003cem\u003e1,2\u003c/em\u003e). These nanostructures produce iridescent colors by the scattering of light from alternating layers of different refractive indices. While a single layer of melanosomes under the keratin cortex is sufficient to produce iridescence (\u003cem\u003e3-6\u003c/em\u003e), additional variation in dimensions and numbers of layers creates additional complexity. As such, melanosomes can vary in size, spacing, and/or shape (being flattened, hollow or both), and can also be organized in layers of different sizes and arrangements, which increases the number of interfaces available for scattering (\u003cem\u003e3,7\u003c/em\u003e) and thereby produces more intense colors. Indeed, empirical research has consistently shown that specific melanosome dimensions, particularly thin, elongated or hollow melanosomes are associated with iridescent coloration (\u003cem\u003e7-9\u003c/em\u003e).\u003c/p\u003e\n\u003cp\u003eWhile the physical mechanisms of iridescent color production have been intensively investigated (\u003cem\u003ee.g., 3,5,6,10-12\u003c/em\u003e), how iridescence evolved remains less clear and is mostly limited to single clades such as ducks, cuckoos and sunbirds (\u003cem\u003ee.g.,7,11-15, but see 16\u003c/em\u003e). Melanosomes that are thin-elongate rods or of specialised shape that are deposited as layers in a keratin feather matrix, were present in avian and non-avian dinosaurs from the Jurassic (\u003cem\u003e17\u003c/em\u003e) and the Cretaceous (\u003cem\u003eMicroraptor\u003c/em\u003e, \u003cem\u003eEoconfuciusornis\u003c/em\u003e, \u003cem\u003eWulong bohaiensis\u003c/em\u003e) (\u003cem\u003e18-20\u003c/em\u003e), suggesting that iridescence was present in the earliest ancestors of birds (\u003cem\u003e16\u003c/em\u003e). However, these studies do not consider the substantial variation in intensity of iridescent coloration that ranges from the subtle sheen of swifts to the brilliant, intense iridescence of peacocks. Nor do they take the topology of coloration, i.e. where iridescence appears on the body, into account, even though this likely influences its function and how it evolves or intensifies (\u003cem\u003e13,21-26\u003c/em\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eClose to 70 years ago, pioneering researchers (\u003cem\u003e3,27\u003c/em\u003e) established a classification scheme with four different degrees of iridescence (weak, moderate, strong, brilliant) (See material and methods for definitions). These categories correspond to colors represented by, for example, swifts (weak), many corvids (moderate), ducks (strong) and hummingbirds (brilliant) (Figure 1A). These groups have been previously used and validated (\u003cem\u003e6,15,28\u003c/em\u003e), and new measurements on 105 specimens confirm that they are repeatable and quantifiable (see below). Here we scored the intensity of iridescence (weak to brilliant) in 71,536 body patches from 8,942 specimens of 5,755 bird species (covering 98% of genera and all families (\u003cem\u003e29\u003c/em\u003e)) to investigate how and when iridescence, and differences in intensity, evolved in birds and extrapolate these results to non-avian dinosaurs.\u003c/p\u003e"},{"header":"Results and Discussion","content":"\u003cp\u003e\u003cstrong\u003eGradual evolution of iridescence from both melanized and non-melanized feather patches.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIridescence was observed in 123 of 193 bird families (64%) (Supplementary data 1,2; Figure 1; Supplementary figure 1-2), substantially expanding the number of families previously (n=85) known to exhibit iridescence (\u003cem\u003e3,6,15,27,\u0026nbsp;\u003c/em\u003ebut see\u003cem\u003e\u0026nbsp;16\u003c/em\u003e, which only included iridescence stronger than weak, i.e. at least moderate). Furthermore, iridescence is present in 1,793 species (30% of species sampled) and 654 genera (30% of genera sample) (Supplementary data 1). Even though iridescence produces some of the most conspicuous colors, it is much more common than previously realized.\u003c/p\u003e\n\u003cp\u003eTo identify transition patterns, we modelled the relative frequencies of gains and losses under five different evolutionary models (Supplementary table 1). These analyses indicated that overall, ordered models perform best (Supplementary table 2-3). More specifically, \u003cem\u003eordered 2\u003c/em\u003e but also \u003cem\u003eordered 1\u003c/em\u003e had the highest AIC values across most datasets. While \u003cem\u003eordered 1\u003c/em\u003e enforces all gains within iridescence classes (i.e. scores \u0026gt; 1) to be gradual (i.e. states cannot jump over adjacent iridescent states), \u003cem\u003eordered 2\u003c/em\u003e is less stringent with a single constraint: brilliant iridescence only evolves out of strong iridescence. This suggests that transitions between different classes of iridescence\u0026mdash;and between gains and losses\u0026mdash;do not occur at equal frequencies. More specifically, they suggest that iridescence, in particular the evolution of brilliant iridescence, is highly gradual. Averaged across all models and patches, rates of evolution (Figure 2) from non-adjacent states (on average, r\u003csub\u003e0-3\u0026rarr;5; male and female, strict and extended\u003c/sub\u003e) to brilliant iridescence are 0 or close to it. Similarly, on average, skipping from weak iridescence to strong iridescence (on average, r\u003csub\u003e2\u0026rarr;4; male\u003c/sub\u003e = 0.004, r\u003csub\u003e2\u0026rarr;4; male_extended\u003c/sub\u003e = 0.008, r\u003csub\u003e2\u0026rarr;4; female\u003c/sub\u003e =0.002, r\u003csub\u003e2\u0026rarr;4; female_extended\u003c/sub\u003e =0.004) is on average three to eight times less likely than strong iridescence evolving out of moderate iridescence (on average, r\u003csub\u003e3\u0026rarr;4; male\u003c/sub\u003e =0.012, r\u003csub\u003e3\u0026rarr;4; female\u003c/sub\u003e =0.016, r\u003csub\u003e3\u0026rarr;4\u003c/sub\u003e \u003csub\u003efemale_extended\u003c/sub\u003e =0.011), but not in the extended male datasets (r\u003csub\u003e3\u0026rarr;4; male_extended\u003c/sub\u003e =0.001). Going from non-iridescence towards moderate (r\u003csub\u003e0,1\u0026rarr;3; male\u003c/sub\u003e \u0026lt; 0.002, r\u003csub\u003e0,1\u0026rarr;3; male_extended\u0026nbsp;\u003c/sub\u003e\u0026lt;0.009, r\u003csub\u003e0,1\u0026rarr;3; female\u0026nbsp;\u003c/sub\u003e\u0026lt;\u003csub\u003e\u0026nbsp;\u003c/sub\u003e0.002, r\u003csub\u003e0,1\u0026rarr;3; female_extended\u0026nbsp;\u003c/sub\u003e\u0026lt;\u003csub\u003e\u0026nbsp;\u003c/sub\u003e0.004) or strong (r\u003csub\u003e0,1\u0026rarr;4; male\u003c/sub\u003e =0, r\u003csub\u003e0,1\u0026rarr;4; male_extended\u003c/sub\u003e=0 , r\u003csub\u003e0,1\u0026rarr;4; female\u003c/sub\u003e = 0, r\u003csub\u003e0,1\u0026rarr;4; female_extended\u003c/sub\u003e = 0) iridescence while skipping weak iridescence is also more unlikely than strong (on average, r\u003csub\u003e3\u0026rarr;4; male\u003c/sub\u003e =0.016, r\u003csub\u003e3\u0026rarr;4; female\u003c/sub\u003e = 0.016, r\u003csub\u003e3\u0026rarr;4\u003c/sub\u003e \u003csub\u003efemale_extended\u003c/sub\u003e =0.011) and moderate (on average, r\u003csub\u003e2\u0026rarr;3; male\u003c/sub\u003e =0.055, r\u003csub\u003e2\u0026rarr;3; male_extended\u003c/sub\u003e =0.094, r\u003csub\u003e2\u0026rarr;3; female\u003c/sub\u003e = 0.054, r\u003csub\u003e2\u0026rarr;3; female\u003c/sub\u003e =0.076,) iridescence evolving out of their adjacent state, except for the extended male dataset (r\u003csub\u003e3\u0026rarr;4; male_extended\u003c/sub\u003e =0.001). Similar, often even stronger, patterns of gradual evolution were observed across datasets, patches and perturbations. For example, while, on average, in the male extended dataset it seems that strong iridescence is less likely to evolve out of moderate iridescence compared to even weaker intensities, this exception is lost in the perturbated dataset, and seems to be driven by high r\u003csub\u003e3\u0026rarr;4\u0026nbsp;\u003c/sub\u003erates in crown, throat and belly. These results strongly suggest that the evolution of iridescence was gradual, with non-ordered models being the best model in 3 (all ARD models) out of 72 tested configurations, in which the rates are still mostly consistent with an ordered model (Supplementary Figure 4-11). Such gradual evolution might reflect the need for novel melanosome structures to produce increasingly intense colors, as has been observed in Starlings (Sturnidae)\u003csup\u003e\u0026nbsp;\u003c/sup\u003e(\u003cem\u003e7\u003c/em\u003e), or further modification of barbules.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;While iridescence is relatively common across the avian phylogeny, it has been lost more often than gained (Figure 2, 3; Supplementary figures 3-11; Supplementary data 3 \u0026ndash; see dryad for files) a pattern previously recovered for Icterids (\u003cem\u003e30\u003c/em\u003e) and Cuckoos (\u003cem\u003e15\u003c/em\u003e) but not for starlings\u003csup\u003e\u0026nbsp;\u003c/sup\u003e(\u003cem\u003e7\u003c/em\u003e). As predicted, iridescence most frequently evolves out of non-organised melanosomes (on average, r\u003csub\u003e1\u0026rarr;2-5; male\u003c/sub\u003e\u0026lt; 0.002; r\u003csub\u003e1\u0026rarr;2-5; male_extended\u003c/sub\u003e \u0026lt; 0.002; r\u003csub\u003e1\u0026rarr;2-5; female\u0026nbsp;\u003c/sub\u003e\u0026lt; 0.007; r\u003csub\u003e1\u0026rarr;2-5; female_extended\u003c/sub\u003e \u0026lt; 0.006). However, while melanin is essential to produce most iridescent feathers, our ancestral state estimations showed that for multiple patches (and consistently across datasets) transition rates for non-melanized (on average, r\u003csub\u003e0\u0026rarr;2-5; male\u003c/sub\u003e\u0026lt;0.002; r\u003csub\u003e0\u0026rarr;2-5; male_extended\u003c/sub\u003e\u0026lt;0.004; r\u003csub\u003e0\u0026rarr;2-5; female\u003c/sub\u003e\u0026lt;0.002; r\u003csub\u003e0\u0026rarr;2-5; female_extened\u003c/sub\u003e\u0026lt;0.004) states are not zero, meaning that the presence of (disorganised) melanin in the ancestral patch is not essential for iridescence to evolve. That is, ordered melanin can evolve out of a feather with or without disordered melanin. Indeed, the existence of non-melanosome-based iridescence in the feathers of manakins and tanagers (\u003cem\u003e31, 32\u003c/em\u003e); but also in other integumentary types such as the beak (\u003cem\u003e33\u003c/em\u003e), are consistent with the results observed here. Alternatively, melanosome-based iridescence might also evolve out of non-melanized barbules. While this occurred many times, one clear example is \u003cem\u003eLeptocoma\u003c/em\u003e where putative ancestral \u003cem\u003eL. minima\u003c/em\u003e, \u003cem\u003eL. brasiliana\u003c/em\u003e and \u003cem\u003eL. zeylonica\u003c/em\u003e had white or carotenoid-bellies, while the derived \u003cem\u003eL. aspasia\u003c/em\u003e and \u003cem\u003eL. calcostetha\u003c/em\u003e have iridescent belies. Previous studies did not find such results (i.e. iridescence evolved out of melanized feathers, e.g., \u003cem\u003e30\u003c/em\u003e)\u003csup\u003e\u0026nbsp;\u003c/sup\u003elikely stemming from the exclusion of non-melanized species in earlier analyses.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOur results thus suggest that other factors in addition to the presence of melanosomes are important for the evolution of iridescent colors, such as the dimensions and elongation of the barbule (\u003cem\u003e10, 15, 30, 34, 35\u003c/em\u003e). Flattening of the barbule provides more surface area for reflection and, when combined with increasing melanosome density, could promote passive self-assembly and reorganization of melanosomes into ordered nanostructures\u003csup\u003e\u0026nbsp;\u003c/sup\u003e(\u003cem\u003e30\u003c/em\u003e). Whether barbule morphology contributes in this way to the likelihood of iridescence evolution is a fascinating question for future research.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFinally, patches do not evolve in isolation, so iridescence in one area may be influenced by the presence of nearby melanized feathers\u003csup\u003e\u0026nbsp;\u003c/sup\u003e(\u003cem\u003e14\u003c/em\u003e). However, the maximum iridescent value across all patches (i.e. the highest score attributed to any patches per species), which eliminates the effect of scoring individual patches by reducing each species to a single maximum score, shows similar rates from non-melanized and melanized feathers to iridescent feathers (e.g., r\u003csub\u003e1\u0026rarr;2-5; male\u003c/sub\u003e=0-0.007; r\u003csub\u003e1\u0026rarr;2-5; male_extened\u003c/sub\u003e=0-0.006; r\u003csub\u003e1\u0026rarr;2-5; female\u003c/sub\u003e=0-0.007; r\u003csub\u003e1\u0026rarr;2-5; female_extened\u003c/sub\u003e=0-0.005 and r\u003csub\u003e0\u0026rarr;2-5; male\u003c/sub\u003e=0-0.005; r\u003csub\u003e0\u0026rarr;2-5; male_extened\u003c/sub\u003e=0-0.003; r\u003csub\u003e0\u0026rarr;2-5; female\u003c/sub\u003e=0.0-0.005; r\u003csub\u003e0\u0026rarr;2-5; female_extened\u003c/sub\u003e=0-0.003). As such, the observed pattern does not seem to be an artifact associated with patch delimitation and is consistent with the evolution of iridescence from non-melanized plumage.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBrilliant iridescence is ancestral to birds\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn addition to providing evolutionary transition rates, ASEs also provide us with probability estimates for different iridescent intensities in different nodes (i.e. ancestors). These ASE analyses (Figure 3-4, Supplementary figure 12, Supplementary data 3) show that the male most recent common ancestor (MRCA) of birds was, with a probability higher than 95%, brilliantly iridescent (extended male dataset), in multiple patches (Figure 3; Supplementary table 4), or at least strong iridescent (male and female datasets). The tendency of the extended dataset to produce higher iridescence scores is potentially linked to an increase in power by including sexually monomorphic iridescent species. Males are generally more iridescent than females, thus while the inclusion of mis-sexed specimens could potentially inflate iridescence intensities in the female extended dataset, this is not the case in the male extended dataset. This makes the extended dataset, with more specimens and no overestimation, the most reliable estimations, with brilliant iridescent in the wings, throat and breast in the MRCA of birds and varying degrees of iridescence in other patches (Figure 4). These results hold in a perturbated datasets where iridescence is artificially, on average, down-scored.\u003c/p\u003e\n\u003cp\u003eGiven that extant paleognaths are non-iridescent, an iridescent MRCA seems counterintuitive, however, fossil paleognaths (Lithornithidae), were likely iridescent, consistent with the findings here (\u003cem\u003e36\u003c/em\u003e). Previous work\u003csup\u003e\u0026nbsp;\u003c/sup\u003e(\u003cem\u003e20\u003c/em\u003e) suggested that iridescence was present in the MRCA of birds, but did not quantify in which patches iridescence occurred, nor the intensity of iridescence. Such brilliant iridescent colors can today be observed in spectacular birds such as hummingbirds and birds-of-paradise and often require specialized melanosomes. Furthermore, as iridescence evolves gradually, brilliant iridescence requires an ancestral species with at least strong iridescence, suggesting that some dinosaurs, had at least strong iridescence, and potentially brilliant iridescence, in some body patches.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMost dinosaur color reconstructions rely on the preservation of melanosomes and the known relationship between melanosome shape and coloration (\u003cem\u003e17, 19, 37-41\u003c/em\u003e). As such, these studies show that dinosaur (and pterosaur) feathers were black, brown and reddish-brown. However, these findings fail to provide a broader picture of color evolution and only reveal a limited subset of color space compared to that present in extant birds\u003csup\u003e\u0026nbsp;\u003c/sup\u003e(\u003cem\u003e42\u003c/em\u003e). \u0026nbsp;Given the broad range of colors in birds and virtually all other tetrapod groups, it is unlikely that dinosaurs had only dull melanin-based colors such as brown and black. Thin and flattened melanosomes with clear signs of layering, previously discovered in dinosaurs from the Jurassic\u003csup\u003e\u0026nbsp;\u003c/sup\u003e(\u003cem\u003e17\u003c/em\u003e) and Cretaceous (\u003cem\u003e18\u003c/em\u003e), suggest the presence of at least weak but potentially more intense iridescence in dinosaurs\u003csup\u003e\u0026nbsp;\u003c/sup\u003e(\u003cem\u003e19\u003c/em\u003e). As such, the results of our ASE align with previously identified iridescence in dinosaurs, including differential distribution of iridescence, e.g. as in \u003cem\u003eWulong bohaiensis\u003c/em\u003e (occurring 30 mya after the split between avian and non-avian dinosaurs, suggesting the presence of this pattern well before this split) (\u003cem\u003e19\u003c/em\u003e), but our results add a quantification of its intensity, with brilliant iridescent colors likely present in the ancestor of birds, and potentially dinosaurs (Figure 4).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWhile these brilliant iridescent colors expand the potential color space in plumages from just that of white, red, and black it is possible that non-avian dinosaur color space was even larger. Putative fossilized carotenoids have not been described and ASE using extant birds strongly suggest that an absence of fossilized carotenoids might correspond to a general absence of carotenoids in dinosaur feathers (\u003cem\u003e43\u003c/em\u003e). As such, structural color might indeed have been a good alternative for generating patterns and color for the purpose of signaling. Nevertheless, other rare, clade-specific pigments (penguins, turacins, turacoverdins and psittacofulvins)\u003csup\u003e\u0026nbsp;\u003c/sup\u003eare found in birds (\u003cem\u003e44-46\u003c/em\u003e), and it is plausible that extinct \u0026ldquo;ghost pigments\u0026rdquo;, i.e. unknown pigments that left no trace in the fossil record, were present in dinosaurs further expanding color space.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDifferent iridescent classes are perceivable\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe usage and validity of quantitative, rather than qualitative, classification of coloration has been discussed extensively before (\u003cem\u003e47-49\u003c/em\u003e). Indeed, a potential caveat of this study is the scoring of iridescent classes itself. Yet, colors can be consistently assigned to distinct categories, even across cultures (\u003cem\u003e50-51\u003c/em\u003e), and qualitative studies have provided valuable insights when quantification is difficult (\u003cem\u003e20, 52, 53\u003c/em\u003e). By revisiting and re-scoring 123 species previously scored by Durrer\u003csup\u003e\u0026nbsp;\u003c/sup\u003e(\u003cem\u003e3\u003c/em\u003e) and Auber (\u003cem\u003e27\u003c/em\u003e), we show that there is a strong correspondence between classifications decades apart, even when accounting for obvious misidentifications (see methods (Supplementary data 4). The similarity between classifications ranges from 78 to 85% depending on the inclusion/exclusion of such misidentifications (Supplementary data 4; Supplementary figure 13). Mismatches almost exclusively occurred between subsequent categories (e.g., weak vs. moderate; strong vs. brilliant).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFurthermore, we used reflectance spectrophotometry (n=105) (Supplementary data 5) to objectively measure color and identify the mechanisms behind these different categories. These analyses that chroma and contrast, two key-features of iridescent reflectance curves\u0026ndash;differ significantly among all categories (Supplementary table 5; Supplementary figure 14) (p-values \u0026lt; 0.007 between all categories). This is expected, given that iridescence produces a distinct, often multi-peak reflectance curve, unlike other color mechanisms. Indeed, chroma and contrast both incorporate the difference between maximum and minimum reflectance, which is expected to be higher in more \u0026ldquo;peaky\u0026rdquo; reflectance spectra where lows correspond to the high absorption by melanin and highs correspond to the selective reflectance. Hue (the observed color) and brightness (overall reflectance), however, are significantly different between some, but not all categories. The absence of a strong signal might indicate that all mechanisms can have overall similar brightness levels (e.g. carotenoids can also be bright), and that some mechanisms (e.g. brilliant iridescence) can produce colors that overlap with both carotenoids and other iridescence intensities. Additional support comes from ordered logistic regressions where we show that precision, recall and F1-score (i.e. the harmonic mean of precision and recall) in a model based on reflectance data where between 2 and 6 times higher than a random model (Supplementary table 6 and 7). Non-perfect scores are potentially attributed to measurement error, or color-metrics visible (in 3D, e.g. by moving the specimen) but not measured using point measures. Indeed, we measured iridescence at one specific angle because we classified iridescence based on the most intense color. Potentially this is also influenced by relative change in color between angles. Nonetheless, classification performance increased from weak to brilliant iridescence, and misclassifications were mostly underestimation, suggesting that our results are conservative. Finally, using a perturbated dataset, where mis-scorings were introduced based on the empirically observed error-rates (Supplementary data 4), we re-performed all analyses and found no difference across analyses (see discussion throughout text).\u003c/p\u003e\n\u003cp\u003eThese results indicate that human eyes can distinguish between different classes of iridescence intensity. Notably, iridescence intensity, largely independent of hue (Supplementary table 5, Supplementary figure 14), is a continuous trait but can be categorized into discrete classes. Indeed, previous research used similar terms (weak, strong, brilliant) (\u003cem\u003e6, 15, 28\u003c/em\u003e)\u003csup\u003e\u0026nbsp;\u003c/sup\u003eand found quantifiable support for these groups (\u003cem\u003e15\u003c/em\u003e).\u003c/p\u003e\n\u003cp\u003eSuch categorization suggests that hue and intensity may represent distinct evolutionary and developmental modules that can evolve independently, potentially optimizing traits such as visibility. While shifting hue to align with spectral sensitivities may enhance species-specific conspicuousness, reducing iridescence intensity could, in turn, lower overall conspicuousness. However, this effect seems to apply primarily to weak-to-strong iridescence transitions, as highly brilliant colors at shorter wavelengths appear to be rare\u0026mdash;though exceptions exist, such as in peacocks and the bird of paradise (\u003cem\u003eLophorina\u003c/em\u003e), which were not included in our measurements but show hues consistent with low wavelengths.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eBirds display some of the most elaborate ornaments in nature, often relying on melanosome-based iridescent colors for visual communication. This study shows that iridescence can evolve directly from non-melanized feathers, without passing through a stage of non-iridescent melanin-based coloration, suggesting that melanin alone is not sufficient and other barbule traits play a crucial role. Even a disorganized layer of melanosomes within a keratin cortex can produce weak iridescence, hinting at a low barrier to the origin of this trait.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFinally, the ancestral presence of iridescence in bird feathers, imply that iridescence may have been widespread in extinct species like dinosaurs and pterosaurs, potentially making the Mesozoic era far more colorful than previously thought.\u0026nbsp;\u003c/p\u003e"},{"header":"Material and Methods","content":"\u003cp\u003e\u003cstrong\u003eSpecimen collection and iridescence assessment\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe scored the presence of iridescence in 71,536 body patches, from 8,942 representative specimens (round skins) of 5,755 species of birds (more than 50% of all species) encompassing all but 36 genera (98% complete) and all families (29). We scanned all specimens available and chose a representative male (n=4263) and female (n=3618) of each species. When sex was unknown, we scored the unsexed specimen (n=1060). When multiple morphs were present, we used the most iridescent morph available. We investigated iridescence intensity (i.e. how conspicuous an iridescent color is), not iridescence per se (i.e. how much do colors change with changing angles). For each species we scored the presence and intensity of iridescence in eight patches corresponding to known feather tracts: crown, mantle, rump, tail, wings, throat, breast and belly. Non-iridescent colors were scored as either melanin-based colors (grey and black colors, as well as non-carotenoid brown colors), or as non-melanin-based colors. Iridescence was further scored by intensity, following Auber (1953) (27) and Durrer (1977) (3), where we scored iridescence as weak when “only close inspection reveals iridescence”, moderate when “the primary impression given by the feathers in situ is black – Rook Corvus frugileges”), strong (“head of Mallard Anas platyrhyncos”) and brilliant (“many Trochilidae”). We created four datasets; All hypotheses were tested for the two sexes (female dataset n=3610; male dataset n=4256); separately, as selective forces can act independently, and even divergently, in different sexes (25, 29-34). Additionally, to analyze datasets containing all species, we conducted analyses using the “extended” dataset, which comprises both male (male_extended and female_extended, n=5742) and female specimens. These datasets were based on the original male and female datasets but included additional female, male and/or unsexed specimens such that each species sampled was present in the extended dataset, even if the target sex was absent.\u003c/p\u003e\n\u003cp\u003eTo verify this apparent subjective classification, we compared our scores with those of Auber and Durrer for 123 species and calculated overlap in score assignment between scorers. Next, we compared our scores with those of Auber and Durrer for 123 species (supplementary figure 13) and found that iridescent scores were highly repeatable (Supplementary discussion, Supplementary figure 13). To further confirm the validity of the Auber and Durrer classification, we compared reflectance spectra of 236 specimens from 105 species with known Auber and Durrer classifications (Supplementary data 3). These were measured at the Museum of Comparative Zoology (supplementary data 4) using an Ocean Optics USB2000 spectrophotometer measuring UV-VIS (300-700nm) (calibrated with a dark and white standard) with an PX2 Light Source. We connected the fibre optic cable with a reflection probe holder that was held at an angle of 90º placed directly on the feather patch, with the probe positioned at 0.5 cm from the sample. As we were measuring iridescent patches, we measured at the angle of maximal reflection. We only measured iridescent patches to compare them to the Auber and Durrer classification. Given that older specimens (\u0026gt;50 years) (54) might have altered colors, we took precautionary measures (55) and excluded specimens that showed physical damage and dusting. Three measurements per patch were taken, negative values were removed and spectra were smoothed using the R package ‘pavo’ (56).\u003c/p\u003e\n\u003cp\u003eTo quantitatively and objectively measure iridescent colors of the different classes proposed by Auber and Durrer, we calculated summary statistics per species and used phylogenetic ANOVA implemented using phytools’s (57) phylANOVA to compare differences in brightness and chroma and saturation between categories (supplementary figure 14, supplementary table 5). We tested multiple variables using pavo: B2 (Mean brightness; Mean relative reflectance over the entire spectral), B3 (Intensity: Maximum relative reflectance), S3 (Chroma: Reflectance over the Rmax +- 50nm range divided by B1, i.e. the sum of the relative reflectance over the entire spectral range), S6 (Contrast: Rmax – Rmin), S10 (Peaky chroma: (Rmax - Rmin)/B2), H2 (Hue: Wavelength at bmaxneg), and finally S9 (Carotenoid chroma: (R700 - R450)/R700) as a control. These showed that there were consistent, and detectable differences across iridescence categories (weak, moderate, strong, brilliant), allowing the subsequent use in the downstream analyses. In addition, to calculate precision, recall and F1 (i.e. the harmonic mean of precision and recall) we performed an ordered logistic regression implemented in MASS (58) with the reflectance data summary parameters as predictors. These results (Supplementary table 6-7) showed that our model outperforms a random model. Nonetheless, we accounted for uncertainty in the analyses below by including perturbated datasets, in which we randomly changed iridescent scores based on the misclassifications between Durrer, Auber, and this study. By using these empirically observed misclassification rates (Supplementary Figure 13), we exclude unlikely misidentifications (e.g. weak for brilliant), but still introduce up to 50% variability (e.g. in weak iridescence where most misidentifications occur).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAncestral state estimations (ASE)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll analyses (here and in the sections below) were run separately for males and females, as well as a male and female “extended” dataset, that included all male/female specimens but were expanded with species for which we either had no information on the sex, or specimens were only available for one sex. All analyses were run on the most recent all-bird phylogeny (59).\u003c/p\u003e\n\u003cp\u003eFor ASE specifically, we estimated the evolution of all eight patches, as well as the maximum degree of iridescence observed across all patches using fitMk in phytools (57). We used fitMk to test five different evolutionary models, the symmetric rates (SYM), the equal rates (ER), the all-rates-different model (ARD), and two ordered models. More specifically, \u003cem\u003eordered 1\u003c/em\u003e is an ARD model that enforces all gains within iridescence classes (i.e. scores \u0026gt; 1) to be gradual (i.e. states cannot jump over adjacent iridescent states), \u003cem\u003eordered 2\u003c/em\u003e is less stringent with a single constraint: brilliant iridescence only evolves out of strong iridescence. We compared models using AIC values. Ancestral states were then estimated using the ancr function, which computes marginal ancestral states. All analyses were rerun on perturbated datasets.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments:\u003c/strong\u003e We thank the EON lab for discussing the results. We also thank Annelore Nackaerts from RMCA, Olivier Pauwels from RBINS, Jeremiah Trimble and Kate Eldridge from MCZ for access to their collections. Finally, this work would have been impossible without the multiple generations of scientists that have collected the birds that were used to quantify iridescence.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e B.A.E.F. (Belgian American Educational Foundation) (M.P.J.N)\u003c/p\u003e\n\u003cp\u003eFWO (Fonds Wetenschappelijk Onderzoek) (M.P.J.N)\u003c/p\u003e\n\u003cp\u003eUGent BOF mandate (M.P.J.N)\u003c/p\u003e\n\u003cp\u003eEuropean Research Council under the European Union’s Horizon 2020 research and innovation program; grant agreement No. 101000504 (M.C.)\u003c/p\u003e\n\u003cp\u003ePortuguese Foundation for Science and Technology (FCT, https://www.fct.pt) research contract CEECINST/00014/2018/CP1512/CT0002 (M.C.).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization: M.P.J.N., R.C.K.B., S.V.E., M.C., L.D.A., M.D.S.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMethodology: M.P.J.N., L.D.A., M.D.S.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eInvestigation: M.P.J.N. with help from G.D.\u003c/p\u003e\n\u003cp\u003eVisualization: M.P.J.N.\u003c/p\u003e\n\u003cp\u003eSupervision: R.C.K.B., S.V.E., M.C., L.D.A., M.D.S.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWriting – original draft: M.P.J.N.\u003c/p\u003e\n\u003cp\u003eWriting – review \u0026amp; editing: all authors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u003c/strong\u003e There are no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData and materials availability:\u003c/strong\u003e All data and code to reproduce the results are available in the main text or the supplementary materials.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cbr\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eG.E., Hill, G. E., K. J., McGraw, (Eds.), \u003cem\u003eBird coloration\u003c/em\u003e (Vols. 1\u0026ndash;2). (2006, Harvard University Press).\u003c/li\u003e\n\u003cli\u003eL., D\u0026rsquo;Alba, M.D., Shawkey, M. D. Melanosomes: Biogenesis, properties, and evolution of an ancient organelle\u003cem\u003e, Physiological Reviews\u003c/em\u003e\u003cstrong\u003e99\u003c/strong\u003e(1), 1-19, (2019).\u003c/li\u003e\n\u003cli\u003eH., Durrer. Schillerfarben der Vogelfeder als Evolutionsproblem. (1977, University of Basel).\u003c/li\u003e\n\u003cli\u003eS. M. Doucet, M. D. Shawkey, G. Hill, R. Montgomerie, Iridescent plumage in satin bowerbirds: structure, mechanisms and nanostructural predictors of individual variation in color, \u003cem\u003eJournal of Experimental Biology\u003c/em\u003e\u003cstrong\u003e209\u003c/strong\u003e, 380-390 (2006).\u003c/li\u003e\n\u003cli\u003eP. Freyer, B. D. Wilts, D. G. Stavenga, Cortex thickness is key for the colors of iridescent starling feather barbules with a single, organized melanosome layer, \u003cem\u003eFrontiers in Ecology and Evolution\u003c/em\u003e\u003cstrong\u003e9\u003c/strong\u003e, 746254 (2021).\u003c/li\u003e\n\u003cli\u003eK. K. Nord\u0026eacute;n, C. M. Eliason, M. C. Stoddard, Evolution of brilliant iridescent feather nanostructures, \u003cem\u003eeLife\u003c/em\u003e\u003cstrong\u003e10\u003c/strong\u003e, e71179 (2021).\u003c/li\u003e\n\u003cli\u003eR. Maia, D. R. Rubenstein, M. D. Shawkey, Key ornamental innovations facilitate diversification in an avian radiation, \u003cem\u003eProceedings of the \u003c/em\u003e\u003cem\u003eNational Academy of Sciences U.S.A.\u003c/em\u003e\u003cstrong\u003e110\u003c/strong\u003e, 10687-10692 (2013).\u003c/li\u003e\n\u003cli\u003eH. Durrer, W. Villiger, Schillerfarben der Trogoniden, J. Ornithol. 107, 1-26 (1966).\u003c/li\u003e\n\u003cli\u003eD. G. Stavenga, H. L. Leertouwer, B. D. Wilts, Magnificent magpie colors by feathers with layers of hollow melanosomes, \u003cem\u003eJournal of Experimental Biology\u003c/em\u003e\u003cstrong\u003e221\u003c/strong\u003e, jeb174656 (2018).\u003c/li\u003e\n\u003cli\u003eR. Maia, R. H. Macedo, M. D. Shawkey, Self-assembly of iridescent barbules through depletion attraction of melanosomes during keratinization, \u003cem\u003eJournal of the Royal Society Interface\u003c/em\u003e\u003cstrong\u003e9\u003c/strong\u003e, 734-743 (2011).\u003c/li\u003e\n\u003cli\u003eC. M. Eliason, R. Maia, J. L. Parra, M. D. Shawkey, Signal evolution and morphological complexity in hummingbirds (Aves: Trochilidae), \u003cem\u003eEvolution\u003c/em\u003e\u003cstrong\u003e74\u003c/strong\u003e, 447-458 (2020).\u003c/li\u003e\n\u003cli\u003eC. M. Eliason, R. Maia, M. D. Shawkey, Modular color evolution in birds facilitated by a complex nanostructure, \u003cem\u003eEvolution\u003c/em\u003e\u003cstrong\u003e69\u003c/strong\u003e, 357-367 (2015).\u003c/li\u003e\n\u003cli\u003eM. P. J. Nicola\u0026iuml;, S. Rogalla, M. Yousefi, R. C. K. Bowie, L. D\u0026apos;Alba, M. D. Shawkey, Ecological, genetic and geographical divergence explain differences in coloration among sunbird species (Nectariniidae), \u003cem\u003eEcology and Evolution\u003c/em\u003e\u003cstrong\u003e14\u003c/strong\u003e, e11427 (2024).\u003c/li\u003e\n\u003cli\u003eM. P. J. Nicola\u0026iuml;, B. Van Hecke, S. Rogalla, G. Debruyn, R. Bowie, N. J. Matzke, S. Hackett, L. D\u0026apos;Alba, M. Shawkey, The evolution of multiple color mechanisms is correlated with diversification in sunbirds (Nectariniidae), \u003cem\u003eSystematic Biology\u003c/em\u003e syae006 (2024).\u003c/li\u003e\n\u003cli\u003eK. K. Nord\u0026eacute;n, C. R. Cooney, F. Babarovic, M. C. Stoddard, Historical contingency influences the diversity of feather nanostructures in cuckoos, \u003cem\u003ebioRxiv\u003c/em\u003e (2023).\u003c/li\u003e\n\u003cli\u003eM. Eliason, M. P. J. Nicola\u0026iuml;, C. Bom, E. Blom, L. D\u0026apos;Alba, M. D. Shawkey, Transitions between color mechanisms affect speciation dynamics and range distributions of birds, \u003cem\u003eNature Ecology and Evolution\u003c/em\u003e\u003cstrong\u003e8\u003c/strong\u003e, 1723-1734 (2024).\u003c/li\u003e\n\u003cli\u003eD. Hu, J. A. Clarke, C. M. Eliason, R. Qiu, Q. Li, M. D. Shawkey, C. Zhao, L. D\u0026apos;Alba, J. Jiang, X. Xing, A bony-crested Jurassic dinosaur with iridescent plumage highlights complexity in early paravian evolution, \u003cem\u003eNature Communications\u003c/em\u003e\u003cstrong\u003e9\u003c/strong\u003e, 217 (2017).\u003c/li\u003e\n\u003cli\u003eQ. Li, K. Gao, Q. Meng, J. A. Clarke, M. D. Shawkey, L. D\u0026apos;Alba, R. Pei, M. Ellison, M. A. Norell, J. Vinther, A new specimen of Microraptor and the evolution of iridescent plumage color, \u003cem\u003eScience\u003c/em\u003e\u003cstrong\u003e335\u003c/strong\u003e, 1215-1219 (2012).\u003c/li\u003e\n\u003cli\u003eY. Pan, Z. Li, M. Wang, T. Zhao, X. Wang, X. Zheng, Unambiguous evidence of brilliant iridescent feather color from hollow melanosomes in an Early Cretaceous bird, \u003cem\u003eNational Science Reviews\u003c/em\u003e\u003cstrong\u003e9\u003c/strong\u003e, nwab227 (2021).\u003c/li\u003e\n\u003cli\u003eD. Croudace, C. Shen, J. L\u0026uuml;, S. M. Brusatte, J. Vinther, Iridescent plumage in a juvenile dromaeosaurid theropod dinosaur, \u003cem\u003eActa Palaeontologica Polonica\u003c/em\u003e\u003cstrong\u003e68\u003c/strong\u003e, 213-225 (2023).\u003c/li\u003e\n\u003cli\u003eS. M. Doucet, M. G. Meadows, Iridescence: A functional perspective, \u003cem\u003eJournal of the Royal Society Interface\u003c/em\u003e\u003cstrong\u003e6\u003c/strong\u003e, S115-S132 (2009).\u003c/li\u003e\n\u003cli\u003eK. Kjernsmo, H. M. Whitney, N. E. Scott-Samuel, J. R. Hall, H. Knowles, L. Talas, I. C. Cuthill, Iridescence as camouflage, \u003cem\u003eCurrent Biology\u003c/em\u003e\u003cstrong\u003e30\u003c/strong\u003e, 551-555 (2019).\u003c/li\u003e\n\u003cli\u003eR. K. Simpson, A. F. Mistakidis, S. M. Doucet, Natural and sexual selection shape the evolution of color and conspicuousness in North American wood-warblers (Parulidae), \u003cem\u003eBiological Journal of the Linnean Society\u003c/em\u003e\u003cstrong\u003e130\u003c/strong\u003e, 89-100 (2020).\u003c/li\u003e\n\u003cli\u003eD. Gomez, M. Th\u0026eacute;ry, Simultaneous crypsis and conspicuousness in color patterns: Comparative analysis of a Neotropical rainforest bird community, \u003cem\u003eAmerican Naturalist\u003c/em\u003e\u003cstrong\u003e169\u003c/strong\u003e, S42-S61 (2007).\u003c/li\u003e\n\u003cli\u003eR. S. Marcondes, R. T. Brumfield, Fifty shades of brown: Macroevolution of plumage brightness in the Furnariida, a large clade of drab neotropical passerines, Evolution 73, 704-719 (2019).\u003c/li\u003e\n\u003cli\u003eF. Babarović, C. R. Cooney, Z. K. Varley, L. O. Nouri, N. J. Nadeau, G. H. Thomas, The effects of ecology and behaviour on the evolution of coloration in Coraciiformes, \u003cem\u003eBehavioral Ecology\u003c/em\u003e\u003cstrong\u003e34\u003c/strong\u003e, 387-398 (2023).\u003c/li\u003e\n\u003cli\u003eL. Auber, The distribution of structural colors and unusual pigments in the class Aves, \u003cem\u003eIbis\u003c/em\u003e\u003cstrong\u003e99\u003c/strong\u003e, 463-476 (1957).\u003c/li\u003e\n\u003cli\u003eD. Stuart-Fox, L. Ospina-Rozo, L. Ng, A. M. Franklin, The paradox of iridescent signals, \u003cem\u003eTrends in Ecology and Evolution\u003c/em\u003e\u003cstrong\u003e36\u003c/strong\u003e, 187-195 (2021).\u003c/li\u003e\n\u003cli\u003eJ. F. Clements et al., The eBird/Clements checklist of birds of the world: v2018 (2018).\u003c/li\u003e\n\u003cli\u003eM. D. Shawkey, M. E. Hauber, L. K. Estep, G. E. Hill, Evolutionary transitions and structural mechanisms of avian plumage coloration in grackles and allies (Icteridae), \u003cem\u003eJournal of the Royal Society Interface\u003c/em\u003e\u003cstrong\u003e3\u003c/strong\u003e, 777-783 (2006).\u003c/li\u003e\n\u003cli\u003eB. Igic, L. D\u0026apos;Alba, M. D. Shawkey, Manakins produce bright iridescent feather colors without melanosomes, \u003cem\u003eJournal of Experimental Biology\u003c/em\u003e\u003cstrong\u003e219\u003c/strong\u003e, 1851-1859 (2016).\u003c/li\u003e\n\u003cli\u003eD. C. Skigin, M. E. Inchaussandague, C. D\u0026apos;Ambrosio, A. Barreira, P. Tubaro, How the observed color of the Swallow Tanager (Tersina viridis) changes with viewing geometry, \u003cem\u003eOptik\u003c/em\u003e\u003cstrong\u003e182\u003c/strong\u003e, 639-646 (2019).\u003c/li\u003e\n\u003cli\u003eM. Nicola\u0026iuml;, G. Debruyn, M. Soenens, M. D. Shawkey, L. D\u0026apos;Alba, Nanoscale millefeuilles produce iridescent bill ornaments in birds, \u003cem\u003ePNAS Nexus\u003c/em\u003e pgae138 (2024).\u003c/li\u003e\n\u003cli\u003eJ.-P. Iskandar, C. M. Eliason, T. Astrop, B. Igic, R. Maia, M. D. Shawkey, Morphological basis of glossy red plumage colors, \u003cem\u003eBiological Journal of the Linnean Society\u003c/em\u003e\u003cstrong\u003e118\u003c/strong\u003e, 106-118 (2016).\u003c/li\u003e\n\u003cli\u003eL. D\u0026apos;Alba, M. Meadows, M. Maia, J. S. Yeo, M. Manceau, M. D. Shawkey, Morphogenesis of iridescent feathers in Anna\u0026apos;s hummingbird Calypte anna, \u003cem\u003eIntegrative and Comparative Biology\u003c/em\u003e icab123 (2021).\u003c/li\u003e\n\u003cli\u003eE.M. Eliason, J.A. Clarke, Cassowary gloss and a novel form of structural color in birds. \u003cem\u003eSci. Adv.\u003c/em\u003e6,eaba0187(2020).\u003c/li\u003e\n\u003cli\u003eQ. Li, G. Keqin, J. Vinther, M. D. Shawkey, J. A. Clarke, L. D\u0026apos;Alba, Q. Meng, D. E. G. Briggs, R. O. Prum, Plumage color patterns of an extinct dinosaur, \u003cem\u003eScience\u003c/em\u003e\u003cstrong\u003e327\u003c/strong\u003e, 1369-1372 (2010).\u003c/li\u003e\n\u003cli\u003eR. Carney, J. Vinther, M. D. Shawkey, L. D\u0026apos;Alba, J. Ackermann, New evidence on the color, ultrastructure, and nature of the isolated Archaeopteryx feather, \u003cem\u003eNature Communications\u003c/em\u003e\u003cstrong\u003e3\u003c/strong\u003e, 637 (2012).\u003c/li\u003e\n\u003cli\u003eJ. Vinther, D. E. G. Briggs, R. O. Prum, V. Saranathan, The color of fossil feathers, \u003cem\u003eBiology Letters\u003c/em\u003e\u003cstrong\u003e4\u003c/strong\u003e, 522-525 (2008).\u003c/li\u003e\n\u003cli\u003eK. K. Nord\u0026eacute;n, J. W. Faber, F. Babarović, T. L. Stubbs, T. Selly, J. D. Schiffbauer, P. Peharec \u0026Scaron;tefanić, G. Mayr, F. M. Smithwick, J. Vinther, Melanosome diversity and convergence in the evolution of iridescent avian feathers-Implications for paleocolor reconstruction, \u003cem\u003eEvolution\u003c/em\u003e\u003cstrong\u003e73\u003c/strong\u003e, 15-27 (2019).\u003c/li\u003e\n\u003cli\u003eCincotta, M. Nicola\u0026iuml;, H. B. N. Campos et al., Pterosaur melanosomes support signalling functions for early feathers, \u003cem\u003eNature\u003c/em\u003e\u003cstrong\u003e604\u003c/strong\u003e, 684-688 (2022).\u003c/li\u003e\n\u003cli\u003eM. C. Stoddard, R. O. Prum, How colorful are birds? Evolution of the avian plumage color gamut, \u003cem\u003eBehavioral Ecology\u003c/em\u003e\u003cstrong\u003e22\u003c/strong\u003e, 1042-1052 (2011).\u003c/li\u003e\n\u003cli\u003eS. N. Davis, J. A. Clarke, Estimating the distribution of carotenoid coloration in skin and integumentary structures of birds and extinct dinosaurs, \u003cem\u003eEvolution\u003c/em\u003e\u003cstrong\u003e76\u003c/strong\u003e, 42-57 (2022).\u003c/li\u003e\n\u003cli\u003eD. B. Thomas, C. M. McGoverin, K. J. McGraw, H. F. James, O. Madden, Vibrational spectroscopic analyses of unique yellow feather pigments (spheniscins) in penguins, \u003cem\u003eJournal of the Royal Society Interface\u003c/em\u003e\u003cstrong\u003e10\u003c/strong\u003e, 20121065 (2013).\u003c/li\u003e\n\u003cli\u003eJ. Van Dyck, Reflectance spectra of plumage areas colored by green feather pigments, \u003cem\u003eAuk\u003c/em\u003e\u003cstrong\u003e109\u003c/strong\u003e, 293-301 (1992).\u003c/li\u003e\n\u003cli\u003eR. Arbore, S. Barbosa, J. Brejcha, Y. Ogawa, Y. Liu, M. P. J. Nicola\u0026iuml;, P. Pereira, S. J. Sabatino, A. Cloutier, E. S. K. Poon, C. I. Marques, P. Andrade, G. Debruyn, S. Afonso, R. Afonso, S. G. Roy, U. Abdu, R. J. Lopes, P. Mojze\u0026scaron;, M. Carneiro, A molecular mechanism for bright color variation in parrots, \u003cem\u003eScience\u003c/em\u003e\u003cstrong\u003e386\u003c/strong\u003e, eadp7710 (2024).\u003c/li\u003e\n\u003cli\u003eJ. A. Endler, On the measurement and classification of color in studies of animal color patterns, \u003cem\u003eBiological Journal of the Linnean Society\u003c/em\u003e\u003cstrong\u003e41\u003c/strong\u003e, 315-352 (1990).\u003c/li\u003e\n\u003cli\u003eY. He, Z. K. Varley, L. O. Nouri et al., Deep learning image segmentation reveals patterns of UV reflectance evolution in passerine birds, \u003cem\u003eNature Communications\u003c/em\u003e\u003cstrong\u003e13\u003c/strong\u003e, 5068 (2022).\u003c/li\u003e\n\u003cli\u003eK. Delhey, M. Valcu, C. Muck, J. Dale, B. Kempenaers, Evolutionary predictors of the specific colors of birds, Proc. Natl. Acad. Sci. U.S.A. 120, e2217692120 (2023).\u003c/li\u003e\n\u003cli\u003eE. Gibson, R. Futrell, J. Jara-Ettinger, K. Mahowald, L. Bergen, S. Ratnasingam, M. Gibson, S. T. Piantadosi, B. R. Conway, Color naming across languages reflects color use, \u003cem\u003eProceedings of the \u003c/em\u003e\u003cem\u003eNational Academy of Sciences U.S.A.\u003c/em\u003e\u003cstrong\u003e114\u003c/strong\u003e, 10785-10790 (2017).\u003c/li\u003e\n\u003cli\u003eN. Seddon, J. A. Tobias, M. Eaton, A. \u0026Ouml;deen, Human vision can provide a valid proxy for avian perception of sexual dichromatism, Auk 127, 283-292 (2010).\u003c/li\u003e\n\u003cli\u003eM. P. J. Nicola\u0026iuml;, M. D. Shawkey, S. Porchetta, L. D\u0026apos;Alba, Exposure to UV radiance predicts repeated evolution of concealed black skin in birds,\u003cem\u003e Nature Communications\u003c/em\u003e\u003cstrong\u003e11\u003c/strong\u003e, 2414 (2020).\u003c/li\u003e\n\u003cli\u003eM. P. J. Nicola\u0026iuml;, R. Vanisterbecq, M. D. Shawkey, L. D\u0026apos;Alba, Back in black: Melanin-rich skin color associated with increased net diversification rates in birds, \u003cem\u003eBiology Letters\u003c/em\u003e\u003cstrong\u003e19\u003c/strong\u003e, 20230304 (2023).\u003c/li\u003e\n\u003cli\u003eJ. K. Armenta, P. O. Dunn, L. A. Whittingham, Effects of specimen age on plumage color, \u003cem\u003eAuk\u003c/em\u003e\u003cstrong\u003e125\u003c/strong\u003e, 803-808 (2008).\u003c/li\u003e\n\u003cli\u003eS. M. Doucet, G. E. Hill, Do museum specimens accurately represent wild birds? A case study of carotenoid, melanin, and structural colors in long\u0026ndash;tailed manakins Chiroxiphia linearis, \u003cem\u003eJournal of Avian Biology\u003c/em\u003e\u003cstrong\u003e40\u003c/strong\u003e, 146-156 (2009).\u003c/li\u003e\n\u003cli\u003eR. Maia, H. Gruson, J. A. Endler, T. E. White, pavo 2: New tools for the spectral and spatial analysis of color in R, \u003cem\u003eMethods in Ecology and Evolution\u003c/em\u003e\u003cstrong\u003e10\u003c/strong\u003e, 1097-1107 (2019).\u003c/li\u003e\n\u003cli\u003eL. J. Revell, phytools 2.0: An updated R ecosystem for phylogenetic comparative methods (and other things), \u003cem\u003ePeerJ\u003c/em\u003e\u003cstrong\u003e12\u003c/strong\u003e, e16505 (2024).\u003c/li\u003e\n\u003cli\u003eW. N. Venables, B. D. Ripley, \u003cem\u003eModern Applied Statistics with S\u003c/em\u003e, 4th ed. (Springer, New York, 2002).\u003c/li\u003e\n\u003cli\u003eE. J. McTavish, J. A. Gerbracht, M. T. Holder, M. J. Iliff, D. Lepage, P. Rasmussen, B. Redelings, L. L. Sanchez Reyes, E. T. Miller, A complete and dynamic tree of birds, \u003cem\u003ebioRxiv\u003c/em\u003e 2024.05.20.595017 (2024).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8020267/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8020267/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Colors of iridescent feathers are some of the most striking in nature, but when and how they evolved remains elusive. Using a dataset of 71,536 body patches from 5,755 bird species, we reconstruct the evolution of iridescence. We find that brilliant iridescence, such as that of peacocks, is likely ancestral in birds, and possibly in dinosaurs, but was lost and regained multiple times over evolutionary history. Transitions from weak to brilliant iridescence were gradual, and iridescence evolved from both melanized and non-melanized patches. These findings reshape our understanding of color evolution in birds and their ancestors.","manuscriptTitle":"The evolution of brilliant iridescence in birds","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-06 08:23:50","doi":"10.21203/rs.3.rs-8020267/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"296aa5d2-006d-4e45-bfdd-de6fbeb2c4bf","owner":[],"postedDate":"November 6th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":57531804,"name":"Biological sciences/Evolution"},{"id":57531805,"name":"Biological sciences/Zoology"}],"tags":[],"updatedAt":"2025-11-10T18:40:11+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-06 08:23:50","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8020267","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8020267","identity":"rs-8020267","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}