INTRODUCTION
Colour is one of the most striking and varied components of visual signals throughout the natural world. The cost of producing vivid colours can be critical for maintaining signal honesty. Colours produced by carotenoid pigments are a text-book example of honest signals of individual quality because carotenoids cannot be synthesized by animals (i.e. must be ingested), and have a range of essential physiological functions (Olson & Owens 1998; Svensson & Wong 2011). Consequently, supply may limit carotenoid intake and there may be trade-offs in allocation of pigments to ornamentation versus other physiological roles (Koch & Hill 2018). The literature on carotenoid-based coloration is dominated by studies of birds; however, the majority of vertebrates (ectothermic vertebrates – fish, reptiles, amphibians) can produce yellow to red skin colours using carotenoids and/or a biochemically distinct class of pigments called pteridines (Bagnara & Matsumoto 2006). Pteridines can be used instead of, or together with carotenoids to produce yellow-red colours and the two pigment classes can frequently be found together in coloured integument (Bagnara & Hadley 1973; Bagnara & Matsumoto 2006). Pteridines are synthesised de novowithin pigment cells from abundant purine molecules (Bracher et al. 1998; Ziegler 2003; Braasch et al. 2007) and have limited antioxidant function (McGraw 2005). This suggests that, in contrast with carotenoids, production costs of pteridine-based colours may be minimal and it is unlikely that supply is limited or that there are trade-offs in allocation. However, few studies have examined pteridine pigments in the context of colour signalling and the physiological costs and evolutionary drivers of variation in pteridine pigments remain largely unknown.
What explains the use of carotenoid or pteridine pigments when both can produce spectrally similar colours? One possibility is that pteridines compensate directly or indirectly for variation in the environmental availability of carotenoids (Grether et al. 1999). Under direct compensation, pteridines directly replace carotenoids with a similar hue (carotenoid-mimicry hypothesis). This is expected to result in a negative correlation between the concentrations of similarly coloured carotenoid and pteridine pigments. Under indirect compensation, pteridines compensate for geographic variation in carotenoids, irrespective of hue. Different species or populations may have different colours, depending on local selective pressures, but use a higher proportion of pteridines when carotenoid availability is low. In this case, we expect a positive association between combined carotenoid concentration and environmental carotenoid availability, and a corresponding negative association with combined pteridine concentration, but we do not necessarily expect correlations between specific carotenoid and pteridine pigments. Whether direct or indirect compensation for environmental carotenoid availability explains the prevalence of pteridine pigments among species has not been tested in any taxonomic group, to our knowledge, due to insufficient data on pigment concentrations.
Within the broad classes of carotenoids and pteridines, specific pigments have different hues, are acquired or metabolised in different ways and therefore have different costs and roles in colour production. Carotenoids are produced by plants and the most dominant carotenoids in angiosperms are yellow xanthophylls such as lutein (Heath et al.2013). Insect herbivores generally sequester carotenoids in proportion to the concentration found in the diet (Heath et al. 2013). Red ketocarotenoids, such as astaxanthin and canthaxanthin are comparatively rare in terrestrial ecosystems (primarily produced by microalgae and yeast), but some animals, including birds and turtles, can metabolically convert dietary yellow carotenoids to red ketocarotenoids (Lopeset al. 2016; Mundy et al. 2016; Twyman et al.2016). Due to the cost of metabolic conversion, or low dietary availability for terrestrial animals, ketocarotenoids are more strongly associated with measures of individual quality and sexual selection than dietary yellow carotenoids, particularly in birds (Weaver et al.2018). Pteridines similarly vary in colour from yellow (e.g. sepiapterin, xanthopterin) to red (e.g. drosopterin, erythropterin, riboflavin) but other pteridines (e.g. pterin, biopterin, isoxanthopterin) found in skin pigment cells (chromatophores) are assumed to be colourless. Colourless pteridines can be found in large quantities within chromatophores, though it is unclear whether or how they may contribute to integument coloration (Bagnara & Matsumoto 2006). The different costs and roles in colour production for different types of carotenoids and pteridines influence expected associations with environmental factors and sexual selection. To understand evolutionary drivers of pigment variation, it is therefore essential to quantify specific metabolites in integument tissue; however, this has only been attempted for carotenoids, and only in some groups of birds (Prumet al. 2012; Friedman et al. 2014b, a; Ligon et al.2016).
Here, using an extensive dataset of concentrations of 11 pigments, we test whether pteridine pigments compensate, directly or indirectly, for environmental availability of carotenoids among 27 species of Australian agamid lizards (186 skin samples, 79 individuals, 28 populations with distinct coloration). Specifically, we use highly accurate liquid chromatography-mass spectrometry to quantify pigment concentrations in skin tissues of agamid lizards (McLean et al. 2017; McLeanet al. 2019). We tested whether pigment concentrations are associated with environmental gradients indicative of carotenoid availability. Since this relationship may depend on the strength of sexual selection, we simultaneously tested for relationships between pigment concentrations and proxies for the strength of sexual selection (sexual dichromatism and sexual size dimorphism). To distinguish directversus indirect compensation, we examined correlations between carotenoid and pteridine pigments with similar hue. Lastly, we evaluated how carotenoid or pteridine concentrations covary with skin colour (hue, saturation, luminance).