How many color metrics do we need? Evaluating how different color-scoring procedures explain carotenoid pigment content in avian bare-part and plumage ornaments

For a variety of technical and conceptual reasons, biologists have come to use several different methods to quantify the colors of animals. However, the relative abilities of these different color-scoring procedures to capture variation in the actual color-generating mechanisms-pigment or structural composition of the integument-have never been tested systematically. Here, we examined which commonly employed color metrics predict carotenoid content of ornaments in three avian species (house finch Carpodacus mexicanus, mallard duck Anas platyrhynchos, and zebra finch Taeniopygia guttata). We used spectrophotometry to measure reflectance spectra from beak and feather tissue, calculated numerous color metrics (e.g., hue, chroma, brightness, principal components, and tetrahedral color space position) from these spectra, and determined carotenoid content at the site of color measurement with high-performance liquid chromatography. We found that several principal component, tristimulus, and avian visual model metrics significantly correlated with carotenoid content of house finch feathers and duck beaks. Carotenoid content of mallard beaks was most closely correlated with brightness and saturation metrics, whereas in house finch feathers, carotenoid concentration was best captured by hue and saturation metrics. According to tristimulus scores and visual models, we found that the ultraviolet portion of the spectrum was not an essential predictor of variation in carotenoid content. Also, visual model chromatic contrasts generally were not significant predictors of carotenoid content, although some achromatic contrasts and tetrahedral color space vector parameters were. Our results indicate that numerous methods, especially tristimulus scores, are suitable for capturing pigment-based color variation in two carotenoid-containing ornaments, and we discuss the merits and shortcomings of these different approaches. In contrast, there were no significant relationships between any color metrics and the carotenoid content of zebra finch beaks, suggesting that other color-generating mechanisms besides carotenoids may contribute to color variability in this species.