On the perception, production and function of blue colouration in animals

Authors

  • Kate D. L. Umbers

    Corresponding author
    1. Research School of Biology, Australian National University, Canberra, Australia
    • Department of Biological Sciences, Macquarie University, Sydney, NSW, Australia
    Search for more papers by this author

  • Editor: Steven Le Comber

Correspondence

Kate Umbers, Department of Biological Sciences, Macquarie University, North Ryde, NSW 2109, Australia.

Email: kate.umbers@mq.edu.au

Abstract

Bright colouration in animals has long attracted the attention of physicists, chemists and biologists. As such, studies on the functions of colours are interdisciplinary, focusing on the mechanisms of colour production and maintenance, the physical and chemical properties of the colour-producing elements, and visual systems and behaviour of potential receivers. Blue colouration has received a large share of research attention and is fascinating for several reasons: blue has been attributed to a very broad range of functions, blue is achieved by a great variety of mechanisms (although their production and maintenance costs are currently unclear), and the blue part of the spectrum (450–490 nm) can be perceived by most taxa. This review explores the breadth of studies that propose a function for blue colouration. In so doing, it discusses the diversity of ways in which blue colours are produced both as pigments and structural colours, and that blue visual pigments are common across a broad range of taxa. This analysis of the current literature emphasizes the importance of multidisciplinary hypothesis testing when attempting to elucidate the function of colours, the need for manipulative over correlative evidence for the function of colours, and, as colour research becomes evermore interdisciplinary, the need for well-defined consistent terminology.

Introduction

Elucidating the functions of colours relies on understanding many different factors: colour production and maintenance, physical and chemical properties of colour-producing elements and visual systems and behaviour of potential senders and receivers. Blue colours are of interest for several reasons, although of course these reasons are not necessarily exclusive to blue colours. Firstly, as well as being attributed a wide range of functions (anti-microbial, thermoregulation, intra- and interspecific signalling), blues (like so many hues predominantly caused by structural colouration) do not fit well into current paradigms of the function of colour because the costs of producing and maintaining them are not clear. Secondly, blues are most commonly the product of structural colours and are produced via a wide variety of optical mechanisms. And thirdly, blue wavelengths can be seen by many taxa and thus have clear potential in signalling (Briscoe & Chittka, 2001). This review presents research on the proposed functions of blue colours and, in so doing, begins with a definition of the colour blue, how it is perceived and the optical mechanisms that produce it.

The perception of blue colours

Light of wavelengths between 300 and 700 nanometres (nm) (ultraviolet to red) are seen as colours. The phenomenon of colour is a product of the way an individual perceives different wavelengths of light and is not an inherent property of the object. It is generally accepted that blues are created by perception of the selective reflectance of light between the wavelengths of 450 and 490 nm (Fox & Vevers, 1960), but because colours are a perceptual phenomenon, the viewer's biology ultimately dictates how these (and all) wavelengths are perceived.

The ability to perceive different wavelengths of light as colours arose early in the evolution of animals and has been lost partially or completely many times (Peichl, Behrmann & Kröger, 2001). Consequently the diversity of visible colours varies between taxa (Hunt et al., 1995; Briscoe & Chittka, 2001; Peichl et al., 2001; Vorobyev et al., 2001; Warrant & Locket, 2004; Dawson, 2006; Frentiu et al., 2007) (Table 1). Some taxa (e.g. most birds and butterflies) perceive colours between, 300 and 700 nm (ultraviolet to red) (Briscoe & Chittka, 2001) others only detect the colours between 400 and 700 nm (blue to red) and most insects perceive wavelengths between 300 and 600 nm (ultraviolet to green) (Briscoe & Chittka, 2001). Many receptors that are optimized for the absorption of light of a particular wavelength often absorb light in neighbouring wavelengths. Colour perception also depends on many variables in the environment such as the background against which objects are seen, the clarity of the air or water, and the amount and colour of the ambient light available (Endler, 1990, 1993). The extent of our current understanding of colour vision in animals includes the physiology of lens and compound eyes, and the receptors contained within, but these are only parts of the cumulative process of colour perception. Beyond receptors, what information travels from the eye to the brain, how it is weighted, and the role of the brain in interpreting colour remains largely unclear (Schnitzer & Meister, 2003).

Table 1. A non-exhaustive list of visual pigments of various taxa showing the wavelengths at which their opsins are maximally sensitive (blue: 450–490 nm)
PhylumClassCommon nameSpecies name

Visual pigment

(nm, λ max)

Reference
AnnelidaPolychaetadeep sea annelids

Vanadis sp

Torrea sp

400, 560

460, 480

Wald & Rayport (1977)
ArthropodaInsectainsects varioussee table 1 Briscoe & Chittka (2001)
ArthropodaMalacostracamantis shrimps various

Marshall et al. (1991)

Porter et al. (2009)

ChordataAgnathalampreyGeotria australis515, 610Collin (2009)
ChordataAvespigeonColumba livia409, 453, 567Kawamura, Blow & Yokoyama (1999)
ChordataAvesblue titParus caeruleus371, 448, 503, 563 (double cones: 565, 563)

Hart et al. (2000a)

Hart et al. (2000b)

ChordataAvesblackbirdTurdus merula373, 454, 504, 557 (double cones: 557, 556)

Hart et al. (2000a)

Hart et al. (2000b)

ChordataAvescut-throat finchesAmadina fasciata370, 447, 500, 563 (double cones: 564, 564)

Hart et al. (2000a)

Hart et al. (2000b)

ChordataAvesGouldian finchErythrura gouldiae370, 440, 500, 562 (double cones: 565, 565)

Hart et al. (2000a)

Hart et al. (2000b)

ChordataAveswhite-headed muniasLonchura maja373, 446, 500, 562 (double cones: 564, 562)

Hart et al. (2000a)

Hart et al. (2000b)

ChordataAvesplum-headed finchesNeochmia modesta373, 442, 500, 565 (double cones: 564, 564)

Hart et al. (2000a)

Hart et al. (2000b)

ChordataCondrichthyesblue-spotted mask rayDasyatis kuhlii477, 500, 553Theiss et al. (2007)
ChordataMammaliatammar wallabyMacropus eugenii424, 530Deeb et al. (2003)
ChordataMammalia  variousJacobs (2009)
ChordataMammaliaaye-ayeDaubentonia madagascariensis406Melin et al. (2012)
ChordataOsteichthyes

West Australian seahorse

zebra-snout sea horse

spotted pipefish

Hippocampus subelongatus

Hippocampus barbouri

Stigmatopora argus

460, 520, 537, 560

430, 460, 520, 537, 560

460, 520, 537, 580

Mosk et al. (2007)
ChordataReptiliacane toadBufo marinus433Sillman (1987)
ChordataReptiliaAmerican chameleonAnolis carolinensis462, 503, 625Kawamura & Yokoyama (1993)
MolluscaCephalopodacuttlefishSepia officinalis492

Brown & Brown (1958)

Bellingham et al. (1998)

Mäthger et al. (2006)

CnidariaCubozoabox jellyfishTripedalia cystophora510O'Connor et al. (2009)

Detection of the blue part of the spectrum is perhaps an ancient shared trait among animals (Cashmore et al., 1999). To see blue, an animal requires a visual pigment that absorbs wavelengths from 450 to 490 nm, as well as an opponent receptor and, obviously, the required pathway to their perceptive unit (brain or equivalent ganglion) (Schnitzer & Meister, 2003). Pigments associated with the absorption (and perception) of blue light are cryptochromes, so named because they eluded researchers for many years (Cashmore et al., 1999). Cones and rods sensitive to blue wavelengths have now been discovered in many taxa. This ubiquity suggests that there may be fundamental fitness benefits in detecting and responding to blue light.

Some taxa, such as butterflies, dragonflies and lampreys, have two visual pigments in cones sensitive to the blue part of the spectrum (Meinertzhagen et al., 1983; Yang & Osorio, 1991; Briscoe & Chittka, 2001; Sison-Mangus et al., 2006; Collin, 2009; Wakakuwa et al., 2010), but the advantage is gained from this duplication is unclear (Yokoyama, 1994; Bradbury & Vehrencamp, 1998). Conversely, in some insects and marine mammals, the capacity for reception of the blue wavelengths in cones has been lost. Peichl et al. (2001) showed that marine mammals from two phylogenetically distant groups (Carnivora and Odontoceti) have secondarily lost their visual pigment for blue. The independent loss of a blue receptor may represent a trade off for greater light sensitivity in deep water, but this explanation is problematic given that sensitivity to blue light is still widespread in other marine taxa (Warrant & Locket, 2004). Also, unlike most other nocturnal animals, aye-ayes Daubentonia madagascariensis have retained the ability to detect the blue/violet part of the spectrum with cones, and express the SWS1 opsin pigment gene (λmax 406 nm) (Melin et al., 2012). Melin et al. (2012) suggest that by retaining this gene, aye-ayes might better target the bright blue arils of a local palm Ravenala madagascariensis in bluish twilight light. Despite the unusual nature of the above examples, the adaptive significance of extra receptors in the first instance, their loss in the second, and their retention in the third has not been examined. Finally, what an animal perceives as blue is likely to in part be affected by its ability to perceive other parts of the spectrum. Questions such as, ‘how is the perception of blue hues affected by the perception of ultraviolet?’ have not been investigated, but are essential to fully understanding colour perception.

The production of blue colours

The colour literature contains a large body of work on the physics and chemistry of colour production and blue colours have received considerable research attention (Goodrich & Reisinger, 1953; Dyck, 1971; Veron, 1973; Rohrlich, 1974; Byers, 1975; Filshie, Day & Mercer, 1975; Kazlauskas et al., 1982; Blanquet & Phelan, 1987; Wilson, 1987; Goda & Fujii, 1995, 1998; Brink & Lee, 1999; Vukusic et al., 2001; Kinoshita, Yoshioka & Kawagoe, 2002; Bulina et al., 2004; Prum et al., 2004; Prum & Torres, 2004; Vukusic & Hooper, 2005; Watanabe et al., 2005; Doucet et al., 2006; Bagnara, Fernandez & Fujii, 2007; Simmonis & Berthier, 2012). This research attention may reflect our curiosity about brilliantly blue-coloured animals and the potential that colour-producing mechanisms have for biomimetic industrial applications. Besides special cases, such as that of male satin bower birds Ptilonorhynchus violaceus who collect natural and artificial blue objects for display in courtship (Borgia, Pruett-Jones & Pruett-Jones, 1985), animals must produce their blue colours or sequester them from other animals. Except for the striking abundance and diversity of bioluminescent marine animals (Widder, 2010) and the firefly Amydetes fanestratus that is bioluminescent at a blue-shifted wavelength (538 nm) (Viviani et al., 2011), colour production mechanisms are classified into two main categories: pigmentary and structural. While this dichotomous classification scheme seems convenient, it is potentially misleading, as it does not well represent the underlying biology of colour because pigments and structures often work in concert (Shawkey, Morehouse & Vukusic, 2009).

Pigmentary blues

Pigments are important directly or indirectly in the production of most colours (Shawkey & Hill, 2006; Amiri & Shaheen, 2012). Pigments can be generally defined as molecules that selectively absorb light at various wavelengths. Those wavelengths of light not absorbed are reflected, and it is these that result in the colour. A blue pigment, therefore, absorbs light at wavelengths across the whole visual range with the least absorption in the blue wavelengths (450–490 nm). Pigmentary molecules can be present in an organism in one of two ways: in an extracellular matrix (living or dead, e.g. feathers) or within a cell. Intracellular pigments are contained within the chromatosomes (pigment-containing organelles) of chromatophores (chromatosome-containing cells). Chromatophores of particular colours are named for their hue [e.g. cyanophores are cells containing blue chromatosomes (Goda & Fujii, 1995)].

Animals' red, orange and yellow colours are often achieved by pigments (e.g. carotenoids), but blue pigments are rare, perhaps because they necessitate more complex chemistry. For example, the blue pigment compound of the bryozoan Rhizostoma pulmo requires a very long and highly polarized chain of single-double alternating carbon bonds (Bulina et al., 2004). When they are present blue pigments are more likely to be found in the extracellular matrix for example in the copepod Pontella fera, the crayfish Procambarus clarkii and the abalone Haliotis discus hannai (Herring, 1965; Cheesman, Lee & Zagalsky, 1967; Milicua et al., 1985). Also, many bird species blue pigments such as biliverdins occur in the extracellular matrix of their eggshells (Kilner, 2006; Stoddard & Prum, 2008).

Goda & Fujii (1995) reported the first and only known cyanophores (true blue chromatophores) in the ectoderm of Synchiropus mandarin fishes. Within the cyanophores, the cyanosomes aggregate and disperse in response to various stimuli probably causing the colour change that occurs in these fish (Goda & Fujii, 1998). It seems unlikely that this is the only incidence of a blue cyanophore in nature and research into potential blue pigments will likely turn up more examples.

Structural blues

Structural colours are those whose wavelengths are reflected as a result of optical interference by nanoscale structures in or on an animal's integument. Ultraviolets, violets, and blues are often structural colours (Bagnara et al., 2007). Examples of body parts on which colour-producing nanostructures occur include the scale on a butterfly's wing (Ghiradella, 1991), the barbule of a bird's feather (Prum, 2003; D'Alba et al., 2011), or the arrangement of fibres or granules embedded in a dermal layer (Filshie et al., 1975; Prum & Torres, 2003, 2004; Prum et al., 2004). In vertebrates, the iridophore is a chromatophore that contains crystalline structures (rather than pigments as in most chromatophores) that give rise to blue colouration (Rohrlich, 1974; Clothier & Lythgoe, 1987; Bagnara et al., 2007). Iridophores are often found in association with yellow pigments to produce green colours and when the yellow pigment is reduced (axanthism) the organism appears blue (Bagnara, Frost & Matsumoto, 1978). Blue colours are produced by a much greater diversity of structures than those found in iridophores. There are several categories of structures that preferentially scatter blue light categorized by their degree of order (Fig. 2). Incoherent and quasi-coherent arrangements are subordered and produce low chroma non-iridescent colours. If ordered, however, structures can produce high chroma colours and iridescent effects (wavelength reflected changes based on the reviewer's angle to the object). The effects mentioned earlier can be caused by a variety of mechanisms and there are number of dimensions at which structures are ordered. Structural colours are often purified by accompanying pigments (notably melanin) that lie underneath surface structures, and absorb non-targeted wavelengths (Shawkey & Hill, 2006). In amelanic phenotypes, therefore, some colours may be muddied, faded or completely lost (Siefferman & Hill, 2005a). Given the intimacy with which structures and pigments work together, it seems that the dichotomous ‘structural’ and ‘pigmentary’ colour classification scheme is convenient, but too simplistic. A more complete classification definition could emphasize that structural colours are those in which the structural element of the colour causes reflection of the dominant wavelength while the pigment acts to purify the reflectance of that wavelength by absorbing light in other wavelengths. Pigmentary colours could more completely be defined as those in which reflectance of the dominant wavelength is caused by the reflective properties of the pigment with the addition that they may be enhanced by the presence of highly reflective structures.

Dynamic blues

Both pigmentary and structural colours may be displayed statically, where the colour is ‘on’ for the whole life of an individual, or change reversibly. Those that take place over days to weeks are morphological colour changes (Gabritschevsky, 1927; Insausti & Casas, 2008). For example, in many birds, plumage colour changes upon the commencement of the mating season (Ralph, 1969). Colour change may also occur over a short time frame (milliseconds to hours) in two ways, via mechanical (conceal/reveal) or physiological colour change (Key & Day, 1954a; Filshie et al., 1975; Umbers, 2011). Mechanical colour changes are those in which animals conceal and reveal a patch of colour. The colour patch itself is static, but by the movement of a wing or limb, the patch of colour is revealed to and concealed from the receiver. As such, to the receiver, part of the sender changes colour. For example blue Morpho butterflies use the iridescent patches on their wings to flash colours on and off depending on their angle to the receiver (Vukusic et al., 2002; Wickham et al., 2006) also, alpine katydids Acripeza reticulata reveal bright blue and red stripes on their interabdomnial membranes when threatened (Fig 1, Umbers, unpubl. data). Many changes to and from blue colouration occur via physiological mechanisms such as intracellular granule migrations (Veron, 1973, 1974; Filshie et al., 1975), but little is known about the function of the resultant colour phases. We expect, however, that the ability to change colour may function in physiological and/or signalling processes (Crook, Baddeley & Osorio, 2002; Stuart-Fox, Moussalli & Whiting, 2007).

Figure 1.

Examples of blue colouration in animals: (a) blue button Porpita porpita Dee Why, Australia; (b) blue racer Caenoplana coerulea Kosciuszko NP, Australia; (c) chameleon grasshopper Kosciuscola tristis Kosciuszko NP, Australia; (d) apline crafish Euastacus sp. Kosciuszko NP, Australia (e) vervet monkey (blue scrotum) Chlorocebus pygerythrus Kruger NP, South Africa; (f) mountain katydid Acripeza reticulata Kosciuszko NP, Australia; (g) common blue damselfly Enallagma ebrium Tiny Marsh, Canada; (h) sea lizard Glaucilla marginata Dee Why, Australia; (i) blue sea star Linckia laevigata Heron Island, Australia; (j) satin bower bird, left: male, right: female Ptilorhynchus violaceus Wollemi NP, Australia; (k) five-lined skink Eumeces fasciatus Florida, USA; (l) blue bottle (Physalia uriculus) Dee Why, Australia (Photos: © Kate Umbers).

Figure 2.

Different types of scattering and their categorization: (a) incoherent order at the nanoscale with variably sized particles resulting in incoherent scatter (Tyndall–Rayleigh scatter) often categorized by the reflectance of ultraviolet and blue wavelengths; (b) quasi-coherent order at the nanoscale where particles (nano-spheres) are roughly the same size and are somewhat ordered, often causes ultraviolet-free blue hues; (c) coherent with nano-spheres or tubes can occur in one-dimensional (one layer of spheres, not depicted), two dimensions (many layers of cylinders, not depicted) or three dimensions (multiple layers of spheres) order; (d) surface grating with coherent one-dimensional order at the nanoscale across a surface; (e) multi-layer reflectors with one-dimensional order at the nano-scale in proximo-distal arrangement within an integument; (f) photonic crystals with three dimensions of order at the nano-scale (can occur in one or two dimensions of order, but not often in insects), can produce opalescent effects. The incoherent and quasi-coherent orders cannot produce iridescence, only the various coherent formations can.

The functions of blue

Blue colours are often expected to have a signalling function because to the human observer, they seem obvious and striking. The likelihood that a given animal's blue colour has a function is based on one of two assumptions. Firstly, the handicap principle (Zahavi, 1975) is often applied to blue colours where it is suggested that blue animals are conspicuous in their environment and that only individuals in the best condition can survive to reproduce. Secondly, if blue colours are difficult to produce, their presence may indicate an individual's inherent quality, for example, structural blues may require developmental precision that may indicate ‘good genes’ or a good developmental environment (Siefferman & Hill, 2005b). That blue (or any colour) has no function, ancestrally had a function, or may be in the process of proliferating through a population, should obviously be considered as null hypotheses.

Blue as an intraspecific signal

Blue as indicator of quality

The purity of an individual's colour is often a product of several factors including the ability to sequester pigments from the environment, nutrition and stability during development, or heredity. If a colour patch reflects the true condition of an individual, it may be an honest signal (Guilford & Dawkins, 1991); however, individuals that have preferred colouration without being good quality may be displaying dishonestly or may be colourful as a result of Fisherian runaway selection (Prum, 2010). Because blue colours often require precise development or expensive pigments, honest signalling hypotheses (Dawkins & Guilford, 1991; Maynard Smith, 1991) are commonly invoked to explain their function.

The role of feather colouration in signalling is well studied. In some species, bright plumage correlates with reproductive success and thus may be an honest signal of an individual's quality (Keyser & Hill, 1999). There are several examples in which blue plumage indicates the quality of a potential mate. In eastern bluebirds Sialia sialis, males with brighter blue and ultraviolet colouration are more successful in winning nest hollows, pair earlier in the season, provision nestlings more often (Siefferman & Hill, 2003, 2005a, 2007) and bright blue female colouration has been linked to a good quality diet and thus may indicate her quality (Siefferman & Hill, 2005a). The amount of blue on the body of male grosbeak Guiraca caerulea correlates with larger body size, lower bacterial load, larger territories with more prey and that they feed their first nestlings more frequently than males with less blue plumage (Keyser & Hill, 1999; Shawkey et al., 2007). Thus, in grosbeaks, we may expect that females should pay attention to the blueness of males. Ballentine & Hill (2003), however, reported that male grosbeak blueness is unlikely to be used by females as a mate-choice cue and that its correlation with large territory and body size indicates a role in intrasexual signalling and male–male competition. Also, in blue tits Cyanistes caeruleus blue and ultraviolet crown colouration is not effected by the nutritional quality of the diet (Peters et al., 2011), but is negatively correlated with the fluctuating asymmetry of feathers (Galvan, 2011); the more asymmetrical the bird, the less blue and ultraviolet the crown.

In some systems, blue is used in signals outside of the body. Blue items may be collected from the environment, such as blue ornaments or blue may be produced by an individual, but expressed as blue eggs. Male satin bowerbirds P. violaceus (Fig. 1) collect blue items from the environment (both natural and artificial) and place them in a bower made of dry vegetation, used in courtship displays (Endler & Day, 2006). Females assess this display separately to the chroma of the male's blue plumage, which correlates with mating success (Endler et al., 2005; Savard, Keagy & Borgia, 2011). The number and attractiveness of the bower ornaments (bower quality) may provide females information about the parasite load of the male that owns the bower (Doucet & Montgomerie, 2002). Males often steal items from the bowers of others and bluer items are more likely to be stolen (Wojcieszek et al., 2006; Wojcieszek, Nicholls & Goldzein, 2007a). Exactly what information about a male is portrayed by his bower is not clear, but constraints on building the most attractive bower may keep the owner honest. It is intriguing to imagine how this system evolved, perhaps blue items exploit a pre-existing bias in females where bluer bowers are more attractive. However, why blue in particular is the favoured colour, is unclear. One suggestion is that blue items are naturally rare in forests (Borgia, Kaatz & Condit, 1987; Hunter & Dwyer, 1997; Wojcieszek et al., 2006; Wojcieszek, Nicholls & Goldizen, 2007b).

Blue eggshell colouration is widespread in birds but its adaptive significance is still elusive (Kilner, 2006; English & Montgomerie, 2011). Three major non-exclusive hypotheses have been invoked to explain why some birds' eggs are blue: sexual signalling, mimicry and crypsis (in low light) (Moreno & Osorio, 2003; Soler et al., 2005). There have also been a variety of other hypotheses put forward including filtration of sunlight, enhancing the physical strength of the shell and warning colouration. Little evidence supports these hypotheses (Moreno & Osorio, 2003); however, it is difficult to know whether researcher bias has emphasized this lack of support. Evidence for the sexual selection hypothesis is founded in that, as an antioxidant, biliverdin is beneficial to developing embryos. Thus, males should pay attention to the antioxidant investment a female has made in her eggs and he should provision young according to which ones she has invested in the most (Navarro et al., 2011). Modelling egg colour with various life history traits of 152 species, Soler et al. (2005) found a positive correlation between bluer eggs and increased polygyny and suggested that females advertise their maternal investment to males via egg colour to entice them to feed her young preferentially. Cassey et al. (2008) considered egg colours in the context of an appropriate avian visual system and found only a weak link between maternal reproductive investment and blue eggshell colouration and thus no support for Soler et al. (2005)'s hypothesis. Navarro et al. (2011), however showed in spotless starlings Sturnus unicolor that egg shell colour intensity and the yolk's carotenoid concentration were positively correlated suggesting that colour may be a useful indicator of female investment. However, the intensity of the blue colour alone may be enhanced by an achromatic component that is more easily detectable in low-light nest environments (Avilés, Soler & Pérez-Contreras, 2006; Avilés et al., 2008) Further, in a test of the effect of eggshell colour on paternal provisioning, English & Montgomerie (2011) found that male American robins Turdus migratorius provisioned young nestlings (3 days old) from vivid blue eggs more than those from pale eggs, but this difference did not hold for older (6 or 9 days old) nestlings. Moreover, in the great reed warbler Acrocephalus arundinaceus, Honza et al. (2011) report no association between the blue-green chroma of egg shells and measures of female quality, and also that males did not adjust their investment (in parasite defence) in relation to egg shell chroma.

In Kilner's (2006) review of bird egg colouration, she reported that blue eggs were unusual among cavity nesters, and more often found in some (not all) species that build exposed nests. Kilner (2006) highlighted that if blue eggs are cryptic in exposed nests this adaptation has only been selectively advantageous in some species. Wegrzyn et al. (2011) argued that in cavity-nesting European starlings Sturnus vulgaris the ultraviolet and blue-green eggshell colour does not reflect female condition, but instead suggest that more intensely blue-green egg colouration makes eggs more easily visible in dark cavities. This is an intriguing hypothesis, but clearly, more empirical evidence is needed. Also, studies should be aware of the age of the eggs measured to avoid any confounding effects of fading (Moreno, Lobato & Morales, 2011).

A classic example of blue colour change as a signal is the diet-dependent foot colouration of the blue-footed booby Sula nebouxii. Velando, Beamonte-Barrientos & Torres (2006) showed that the intensity of the blue of a male's feet is a strong indication of his current condition, with the foot colour of nutrient-deprived males fading in less than two days. They also showed that maternal investment reduced when the feet of a male were experimentally dulled using cosmetics (Velando et al. 2006). These results indicate that females adjust their behaviour according to the foot colour of their mate and thus that females receive information on a male's recent foraging success by assessing foot colour. Even though, foot colour fades, it is likely to be a good indicator of recent foraging success and in older birds, an indication of their levels of oxidative stress (Torres & Velando, 2007).

Individual quality may be signalled by blue in inveretebrates. The evidence is sparse, but two examples that involve colour change have emerged. In the damselfly, Calopteryx maculata males with abdomens that are more blue than green are in better condition (Fitzstephens & Getty, 2000). Males that are better foragers increase their girth and in so doing the lamellae (microscopic ridges) in the epicuticle responsible for their blue-green colour are pushed closer together. This results in bluer-looking males because as the ridges get closer together, the shorter (bluer) wavelengths are preferentially reflected. Poorer-condition males, however, look green because the ridges are further apart (Fitzstephens & Getty, 2000). This colour change correlates with the territorial status of a male, but whether blueness translates into fitness benefit via female preference or male–male competition is not yet clear (Fitzstephens & Getty, 2000). Also, recently, Barnard et al. (2012) reported on a blue streak on the anterio–dorsal part of the carapace of sexually mature mud fiddler crabs Uca pugnax, They observed that the streak became darker in colour with decreased ambient light, but did not change with temperature and suggest that its reflectance or rate of change may encode information useful in courtship (Barnard et al., 2012).

Blue for sex identification

In most gonochorist species, there are fitness advantages in displaying one's sex [notable exceptions include: beta male cuttlefish masquerading as females (Hanlon et al., 2005) and andromorphic female dragonflies (Forbes, Richardson & Baker, 1995)]. Some studies assess whether species use colour as a sex cue through manipulative behavioural assays. For example, in many Odonata, a proportion of females don bluer, male colouration (Fincke, 1994; Van Gossum, Stoks & De Bruyn, 2001; Iserbyt et al., 2009) While some studies have found support for the hypothesis that andromorph females endure less harassment by males (Cordero, Carbone & Utzeri, 1998; Van Gossum et al., 2001) or may actually be mimicking males (Robertson, 1985), others have shown that males can learn to recognize andromorphs as females (Miller & Fincke, 1999). Cooper & Burns (1987) found that the blue venter of fence lizards Sceloporus undulatus is used by males to recognize the sex of conspecifics. When presented with females that were painted with male colours, male fence lizards displayed aggression. When presented with males painted with female colours, male fence lizards displayed courtship behaviours. How females react to painted males in this species would be of great interest to determine if colour is used in recognition by both sexes. Also, testing for further functions may reveal that this colour conveys multiple signals, not only sex but something about the quality of the individual.

Male Balkan moor frogs Rana arvalis wolterstorffi change colour from brown to blue and ultraviolet during the mating season (Ries et al., 2008; Hettyey et al., 2009). Ries et al. (2008) suggest that this is so male frogs can ensure they are recognized as such during scramble competition. However, Sheldon et al. (2003) propose that blue male colouration signals genetic quality that helps tadpoles avoid predation. Hettyey et al. (2009) found that the bluest of the small males enjoy greater mating success while blueness of the larger males does not predict mating success. They also reported that bluer individuals had higher body temperatures, but the mechanism of colour change and how it relates to body temperature is unknown. If selectively advantageous, it is unclear as to why males turn blue for only a brief period rather than maintaining their blue and ultraviolet colouration all the time and perhaps suggests a trade off with crypsis. Investigating costs of maintaining their blue colour may be the key to understanding the function of this colour change.

Signalling sex may be particularly important in sequentially hermaphroditic species such as the western Achoerodus gouldii and eastern blue gropers A. viridis and the blue-throated wrasse Notolabrus tetricus. In these species, females turn blue as they become male through a shift in the biliverdin (a blue pigment) concentration in their blood (Gagnon, 2006; Coulson, Hesp & Potter, 2009). If a male is removed from the population the largest female will change sex and in doing so, change the colour to blue (Coulson et al., 2009). The cues for this change and how it affects the behaviour of conspecifics, however, remain unexplored.

Sex identification seems to be given as the function of colouration when a study yields no evidence to support sexual signalling. In this way, sex identification is used like a null hypothesis or default explanation for sexual dichromatism. If the function of colouration is sex identification, it may only be a small evolutionary step away from providing more information than just sex such as information about the individual's quality. Variation in such signals could be co-opted as indicators of quality for preference or in aggressive interactions.

Blue as an interspecific signal

Aposematism

Aposematic colouration is commonly known as warning colouration (Lindström et al., 1999), whereby individuals use bright colours to warn predators that they are distasteful or toxic and therefore should not be eaten or attacked. When warning colours of diverse taxa converge, the species are Müllerian mimics (Merrill & Jiggins, 2009). Alternatively, Batesian mimics cheat by being palatable, but falsely displaying aposematic-like colouration (Rowland et al., 2007). Several studies have reported on the use of blue colour for aposematism, some of which show that blue colouration is used to deter or deflect predators, while others found no evidence to support it.

Some species use aposematic colouration honestly, that is they are brightly coloured and are in fact toxic (Ritland, 1991). The poison dart frogs are a classic example of aposematic colouration (Hoffman & Blouin, 2000). The ‘blue jeans’ strawberry poison dart frog Oophaga pumilio has highly toxic skin and is known for its blue and red display in which each colour is likely to enhance the other because they contrast strongly (Saporito et al., 2007). Saporito et al. (2007) show that red and blue frog models were half as likely to be attacked than brown models. Further experiments could include more model variants to tease part the relative contributions of the colours in the signal. Similarly, blue poison dart frogs Dendrobates azureus are known to have toxic skin (Brodie, Jr. & Tumbarello, 1978), but there has been no empirical study to date testing whether its blue is an aposematic signal. In fact, despite how often blue colours are suggested to be aposematic [e.g. nudibranch – Nembrotha kubaryana (Karuso & Scheuer, 2002), blue-tongued lizard – Tiquila scincoides (Wilsdon, 2009), blue ringed octopus – Hapalochlaena maculosa (Williams, 2010), mountain katydid – Acripeza reticulata (Rentz, 1996)], studies have rarely directly tested the hypothesis.

Redirecting predators

Blue may be used to direct predators to attack dispensable parts of the body (e.g. tail autotomy in skinks) (Cooper & Vitt, 1985). Juvenile American five-lined skinks Plestiodon fasciatus have a distinctive blue tail (Fig. 1), while the adults are cryptically coloured. Clark & Hall (1970) refuted this hypothesis in a study where they conducted behavioural assays and showed that adult male P. fasciatus were less likely to attack a juvenile conspecific if it had a blue tail than if it did not. As such, they suggested that instead of redirecting predators, the blue tail colour enables aggressive adult males to differentiate between other adult males (potential rivals) and juvenile males (not rivals) thus redirecting males to real rivals and reducing infanticide (Clark & Hall, 1970). However, this assertion was refuted by Cooper & Vitt (1985) as they found that adult males readily eat hatchlings with blue tails and thus the primary function of the blue may not be important in intraspecifc signalling after all. Juvenile Acanthodactylus lizards also sport blue tails. Unlike Clark & Hall (1970), Hawlena (2009) suggested that Acanthodactylus use bright blue colouration to direct the attention of predators to their tail. Hawlena (2009) showed that bright blue tail colouration persists in juveniles because their increased activity levels negate any advantages of cryptic colouration, while more sedentary adult Acanthodactylus take advantage of non-blue cryptic colouration. The conflicting results from these studies highlight the need for more empirical data on bright blue juvenile colouration.

Crypsis

Here, crypsis is defined as colouration or morphology that makes detection of an animal more difficult (Stevens & Merilaita, 2009). Crypsis is opposed to mimicry in that mimics actively send deceptive signals (I am a twig, not a phasmid) whereas in crypsis, animals aim to remain undetected (I am not here at all) (Starrett, 1993). There is little evidence for the role of blue colours in crypsis. Macedonia et al. (2009) provide the only evidence of blue colouration being used for crypsis, in Dickerson's collared lizard Crotaphytus dickersonae. They show that in the coastal species, the blue colour of males is more similar to that of the nearby ocean than that of the blue males in the inland sister species. They concede that serpentine and avian predators may not regularly encounter C. dickersonae from angles at which they are framed by the sea, but suggest that it may happen often enough to select for a shade of blue closer to that of the sea than the blue of their inland counterparts (Macedonia et al., 2009). Since terrestrial habitats are predominantly green and brown, it is perhaps not surprising that blue colours are not commonly tested for roles in crypsis. However, it seems that the potential function of blue colouration in crypsis in aquatic, especially marine, habitats should be tested in the future.

Non-signalling functions

For some animals, it is unlikely that their blue colour functions as a signal. However, even when colours are found to be important in signalling, non-signalling functions are rarely tested. Mentioned later are studies that infer a function of blue colouration other than for signalling. These include thermoregulation, antimicrobial activity and protection against solar radiation.

Turning blue for thermoregulation

In the chameleon grasshopper Kosciuscola tristis and dragonflies and damselflies of several genera (e.g. Austrolestes, Diphlebia and Aeshna), males rapidly turn bright blue when their bodies warm up (e.g. K. tristis: body temperature > 25°C) (Key & Day, 1954a,b; Veron, 1974; Sternberg, 1996) (Fig. 1). In these systems, some individuals (often the males) change colour from dark brown to blue while others (usually the females) remain dark in colour. This change in colour is achieved via an intracellular granule migration (Filshie et al., 1975). Key & Day (1954a) and Veron (1974) suggests that the dark phase enhances thermoregulation through maximal absorption across the spectrum. While this does not directly address the function of the blue phase, it implies that the bright colour may afford some protection from overheating (Sternberg, 1995). Veron (1974) and Umbers (2011) show that for dragonflies and grasshoppers, animals in the dark and the blue phases, respectively, heat up at similar rates, and that the blue colouration is unlikely to provide a biologically relevant temperature difference to the black (Umbers, Herberstein & Madin, in press; Veron, 1974). Further, given the sexual dichromatism in many of these species, it seems likely that this colouration has a role in sexual selection, perhaps in displaying body temperature. To test this hypothesis data on the effect of temperature on physiological processes could be coupled with behavioural experiments on male temperature preferences and female preference for male hue. More broadly, studies that focus on the trait in the Odonata could provide insight into why it has repeatedly evolved.

Blue as protection against solar radiation

There is a great diversity of blue organisms found in hauls that skim the top 10 cm of the open ocean, the so-called blue layer (Mollusca, Arthropoda, Cnidaria) (Fig. 1). Herring (1965) reports a blue pigment from the blue layer copepod P. fera. He proposed that P. fera uses its blue pigment to protect its DNA from intense blue wavelengths, but this hypothesis has never been tested. Reflection in the ultraviolet would also provide an organism important protection, but it is unknown whether animals in the blue layer reflect in the ultraviolet. A study that measures the spectral reflectance of blue layer animals would provide interesting insight to the selective pressures that may have lead to their convergence on the colour blue.

Antimicrobial function of blue

Some brightly coloured colonial bryozoans have blue compounds within their tissues such as the blue tetrapyrrol pigment found in Bugula denata (Matsunaga, Fusetani & Hashimoto, 1996). This pigment has antimicrobial properties against both Gram-positive and Gram-negative bacteria, but the source of the blue pigment (whether self-generated or sequestered) is unknown. Similarly an unidentified ascidian (Chordata: Urochordata) from the west coast of Australia also possesses a blue pigment that is likely to have an antimicrobial function. Kazlauskas et al. (1982) showed that the blue pigment from the west Australian ascidian has ‘strong biological activity’. No further description was given and the pigment's function was not tested directly, but such pigments could be indicative of physiological processes such as a by-product of a physiological function.

Conclusions and outlook

Many hypotheses have been invoked to explain animal colouration. These are broadly categorized as either signalling or non-signalling functions. Blue colours have great potential to function in non-signalling roles such as aiding in physiological processes like thermoregulation and protection from harmful solar radiation. However, there is, extremely limited evidence that blue colours function in this way. It is unclear whether researchers are testing these hypotheses, but arriving at null results that are subsequently not published or whether these hypotheses are rarely tested.

Many studies have tested the role of blue colours in signalling and some classic examples of signalling have arisen from these studies (e.g. blue-footed boobies, stain bower birds). But because most taxa can perceive blue, blues may be useful as signals in many species or as broadcast signals to multiple receivers. We may predict for example that a blue signal could simultaneously attract mates and protect from overheating by reflect high energy blue solar irradiance. Because blue colours have the potential to function in several ways simultaneously, multiple hypotheses must be tested if the function of a colour patch is to be understood. To fully extract the function of a colour as a signal, we must understand the mechanism of colour production, what physiological processes a colour's reflectance may influence, the context in which the signal is sent, the visual capacity of the receiver, the light in the environment in which the colour is displayed and how the behavioural response of the receiver changes when the colour shifts.

Clearly, elucidating the function of colouration is a multifaceted problem. Colour research in different disciplines (such as chemistry, physics and biology) has progressed for many years, sometimes in parallel with, but often in isolation from one another. Recently, evolutionary biologists have begun to take a more multidisciplinary approach to research on the function of colours by incorporating the natural history and behaviour of a species with the cellular biology, physical and chemical nature of colour production (e.g. gene regulation to protein morphology) (Siefferman & Hill, 2003; Shawkey & Hill, 2006; Kemp & Rutowski, 2007). Continuing to develop interdisciplinary approaches will enrich the study of animal colouration and lead to the development of novel hypotheses on the evolution of the functions of colour and the ability to test them in new ways.

Acknowledgements

Thanks to Marie E Herberstein, Greg I Holwell, Ainsley E Seago, Anne C Gaskett and Darrell J Kemp for insightful discussion and feedback on earlier versions of the paper and thanks to Ainsley E Seago and Tom D Schultz for helpful discussion about the production of structural colours.

Ancillary

Advertisement