Avian Color Vision and Coloration: Multidisciplinary Evolutionary Biology

A fundamental issue in biology is explaining the diversity of coloration found in nature. Birds provide some of the best-studied examples of the evolution and causes of color variation and some of the most arresting color displays in the natural world. They possess perhaps the most richly endowed visual system of any vertebrate, including UV-A sensitivity and tetrachromatic color vision over the 300–700-nm waveband. Birds provide model systems for the multidisciplinary study of animal coloration and color vision. Recent advances in understanding avian coloration and color vision are due to recognition that birds see colors in a different way than humans do and to the ready availability of small spectrometers. We summarize the state of the current field, recent trends, and likely future directions. A fundamental issue in evolutionary biology is explaining the extraordinary diversity of coloration found in nature. The topic includes the origin and maintenance of sexually selected and naturally selected coloration and textbook examples of coevolutionary processes such as mimicry, warning coloration, camouflage, pollination, seed dispersal, and predator-prey and host-parasite interactions. It has engaged the minds of some of the most eminent evolutionary biologists—from Darwin, Wallace, and Gould to Cott, Ford, Fisher, Tinbergen, and Hamilton. Understanding animal coloration is invariably aided by knowledge of the visual abilities of the animals that have evolved to receive the color signals. But while much of the coloration that demands explanation has been apparent to evolutionary biologists since the 1850s, this is not the case with visual abilities. Color vision in probably the best-studied * Corresponding author; e-mail: [email protected]. Am. Nat. 2007. Vol. 169, pp. S1–S6. 2007 by The University of Chicago. 0003-0147/2007/1690S1-41833$15.00. All rights reserved. invertebrate, the honeybee, was not proved until von Frisch’s (1914) work, although Lubbock (1888) demonstrated color vision in Daphnia and thought that honeybees associated food and color. In humans, trichromacy was first hypothesized in the eighteenth century (Mollon 1989; Kelber et al. 2003), but it was also not until the 1980s that cone spectral sensitivities were first directly measured (Bowmaker and Dartnall 1980). Major patterns in mammalian color vision were not apparent until the 1980s, with the publication of Jacobs’s (1981) book Comparative Color Vision. Over the past two decades, extensive comparative studies have revealed major trends in color vision of vertebrates and invertebrates (reviewed in Goldsmith 1990; Jacobs 1993; Bennett and Cuthill 1994; Cuthill et al. 2000; Briscoe and Chittka 2001; Hart 2001; Kelber et al. 2003; Osorio and Vorobyev 2005; Hart and Hunt 2007). Consequently, accounts of the evolution of the diversity of animal color vision are relatively recent. Key issues here concern the bases for spectral sensitivity functions, the tuning of spectral photoreceptors, and why some taxa are dichromatic (e.g., many mammals), others trichromatic (e.g., humans and bees), and others probably tetrachromatic (e.g., many if not most birds; Kelber et al. 2003). A further puzzle is provided by the finding that while there is much diversity among major taxonomic groups, there is rather little variation in receptor spectral sensitivities within terrestrial taxa such as birds (Hart 2001), Hymenoptera (Briscoe and Chittka 2001), and Old World primates (Kelber et al. 2003). In the past decade, a fundamental change has occurred in how color is measured, considered, and analyzed in evolutionary and ecological studies, particularly of birds. Previously, the implicit assumption was made in virtually all studies of avian coloration that birds saw colors in much the same way as humans (Bennett et al. 1994). A publication in American Naturalist suggested that this approach was flawed (Bennett et al. 1994); the main evidence was that because of the work of Burkhardt, Goldsmith, and colleagues (reviewed in Bennett and Cuthill 1994; Bennett et al. 1994), many birds appeared to be UV sensitive and possibly tetrachromatic; thus, birds appeared not only to S2 The American Naturalist be sensitive to a range of wavelengths to which humans were blind but also to see a range of colors that humans could not perceive. A fundamental change in approach was suggested, including use of spectrometers sensitive to the UV-A and encompassing the entire bird visible waveband (ca. 300–700 nm), in order to assess coloration objectively (Bennett et al. 1994). It is surprising that it took so long for practitioners of color measurement in ecology, behavior, and evolutionary biology to incorporate knowledge of their animals’ spectral range and color vision. In studies of honeybees and insects, UV sensitivity and UV cues were widely considered (e.g., Kevan 1978; Silberglied 1979; Chittka and Menzel 1992). And Lythgoe (1979), Endler (1978, 1990), Burkhardt (1982, 1989), Burkhardt and Finger (1991), and visual physiologists had advocated taking account of an animal’s vision in understanding color signaling, although apart from Burkhardt, the UV-A sensitivity of birds (and their probable tetrachromacy) seems to have been overlooked by these authors. Moreover, tuning of sensory capabilities to relevant sensory information is a fundamental principle of sensory biology (Endler 1978, 1990; Lythgoe 1979; Bradbury and Vehrencamp 1998), and sensory drive and sensory exploitation hypotheses (Ryan 1990; Endler 1992; Basolo and Endler 1995) explicitly hypothesize that sensory capabilities that have evolved for one purpose may be evolutionarily co-opted for other purposes. Ultraviolet vision in birds was known from the early 1970s, having first been demonstrated in hummingbirds by Huth and Burkhardt (1972) and in pigeons by Wright (1972), using behavioral methods (e.g., operant conditioning). So why was it being ignored by practitioners of color measurement in evolutionary and behavioral ecology? There were several reasons. One was the misunderstanding that color as perceived by humans represents an objective reality (Bennett and Cuthill 1994; Bennett et al. 1994). Another was that practitioners of avian color measurement were either unaware of Burkhardt’s and Wright’s findings or did not want to deal with the potential complications arising from them (Bennett and Cuthill 1994; Bennett et al. 1994). A common view in behavioral ecology at the time was that to understand function one did not need to understand the underlying mechanisms (Krebs and Davies 1987). Finally, spectrometers were expensive, cumbersome, and time-consuming to operate, and only in the late 1980s did the first models begin to incorporate the UV waveband in single 300–700-nm scans (e.g., Burkhardt 1989; Burkhardt and Finger 1991). Several factors contributed to the change in approach during the past decade. Increasing numbers of birds and other vertebrates were shown to be sensitive to the UV-A waveband (320–400 nm; e.g., Burkhardt and Maier 1989; Jacobs et al. 1991; Jacobs 1992; Bennett et al. 1996; Bowmaker 1998; Hart et al. 1998, 1999, 2000; Wilkie et al. 1998; Losey et al. 1999; Hunt et al. 2001). Experiments showed that UV information was consistently used in avian mate choice and foraging tasks (e.g., Bennett et al. 1996, 1997; Hunt et al. 1997; Johnsen et al. 1998, 2003; Smith et al. 2002b). Miniaturization and price reductions of spectrometers meant reflectance and radiance spectra over the 300–700-nm range could be readily gathered. Evidence accumulated showing that there was interesting variability in UV reflectances from plumage and skin that had to be explained (Burkhardt 1989; Endler 1990; Burkhardt and Finger 1991; Andersson 1996; Endler and Théry 1996; Bennett et al. 1997; Andersson et al. 1998; Hunt et al. 1998; Prum et al. 1998, 1999; Cuthill et al. 1999; Sheldon et al. 1999). Thus, it was increasingly recognized that it was prudent and practical to consider the UV-A waveband and the entire avian visible range rather than to rely on human color vision or standards based thereon (Bennett and Cuthill 1994; Bennett et al. 1994) when assessing animal coloration. This special issue focuses on birds as a model system. Contributions have been based around presentations at the Second European Conference of Avian Colour Vision and Coloration, in Paris, which we organized with Susana Santos and which brought together many of the leading researchers in avian color vision and coloration. With 12 articles from scientists in Europe, North America, and Australia, it illustrates numerous current approaches and opportunities for application to nonavian taxa. Initial articles concentrate on mechanisms of color vision, with later ones on coloration. Contributions reflect much of the multidisciplinary diversity of research in the field and include retinal physiology and molecular biology of photopigments, psychophysics, and learning rules, colorimetry of feathers, fleshy ornaments and fruits, mechanisms of color production in feathers, and the role of bacteria and possible cosmetics in modifying plumage coloration. All main feather types are included (e.g., structural iridescent, structural blues/UV, pigments). And there are both crossspecies studies and detailed investigations of single species, including the estimation of fitness components.