Discrimination of signal carotenoid content using multidimensional chromatic information

Discrimination of signal carotenoid content using multidimensional chromatic information Abstract Red, orange, and yellow carotenoid-based ornaments, which are widely used as sexual signals in many birds, fish, and reptiles, are known to exhibit multidimensional chromatic variation as a result of both the concentration and relative proportions of different constituent carotenoids with differing spectral properties. This is thought to reflect intrinsic variation in signaler quality, making it a useful basis for female choice. However, whether females are able to discriminate relevant variation in carotenoid concentration and/or composition independently of each other, and of other phenotypic or behavior traits, and if so, how this mediates their choice, is poorly understood. Here, female 3-spined sticklebacks (Gasterosteus aculeatus) were presented with computer-animated courting males that varied exclusively in the appearance of their carotenoid-based coloration; specifically, each male’s signal provided a perceptual match for carotenoid coloration expressed by live males with known underlying carotenoid content, thereby providing a biologically-relevant signal while precluding confounding traits influencing female choice. Females were able to discriminate between prospective mates solely on the basis of perceived variation in the allocation of carotenoids to males’ sexual signals, and exhibited a strong preference for males with coloration indicative of higher concentrations of carotenoids in their signal, rather than in response to perceived variation in the relative proportion of constituent carotenoids. This has important implications for our understanding of male signaling strategies and the information content of carotenoid-based sexual signals. INTRODUCTION Color-based signals are one of the most widely used models in the study of sexual selection, and many studies have shown how individual variation in signal color can affect reproductive success through mate choice and intrasexual competition (Hill and McGraw 2006b; Kemp and Grether 2014). In these studies, color patches are typically considered either as single traits or as individual components of complex (multicomponent or multimodal) signals, and it is often overlooked that they themselves may be perceived as multidimensional signals, with different signal dimensions limited by specific constraints and subject to specific selection pressures (Candolin 2003; Grether et al. 2004). One good example of this is carotenoid-based coloration, which is responsible for much of the yellow, orange, and red regions of coloration present on reptiles, fish, and birds, and is known to exhibit multidimensional chromatic variation (Grether et al. 2004; Rowe et al. 2004; Pike et al. 2011; Romero-Diaz et al. 2013) as a result of variation in both the concentration and identity of the constituent carotenoids, which can differ in their spectral properties (Britton et al. 2004). Pigmented skin and feathers, for example, typically contain several different carotenoid pigments, and variation in the relative concentration of each can markedly affect the color produced (Stradi et al. 1998). Because vertebrates are incapable of producing carotenoid pigments de novo, dietary carotenoids are often used directly, and unchanged, in signaling (Olson and Owens 1998). However, in most species, spectrally distinct forms can be derived from the metabolic transformation of ingested carotenoids (Goodwin 1984; Brush 1990; Stradi et al. 1998; Wedekind et al. 1998; McGraw et al. 2002; Pike et al. 2011), potentially allowing animals control over the appearance of carotenoid-based colors and providing a mechanistic basis for the substantial variation observed both between and within species (Hill and McGraw 2006a). Most studies of carotenoid signaling implicitly assume that a receiver’s decisions are mediated ultimately by this underlying variation in carotenoid allocation, inferred through variation in the color produced. However, while we have a good understanding of the observed variation in carotenoid-based signal expression (Hill and McGraw 2006b), and of the biochemical basis for this variation (Britton et al. 2004), we have a comparatively poor understanding of whether the underlying patterns of carotenoid allocation are actually detectable by receivers and, if they are, how these perceived differences influence behavior. One of the reasons for this ambiguity is due to the inherent difficulty in manipulating different dimensions of the same signal independently in live animals without inadvertently affecting the expression of other cues. For example, dietary manipulation of carotenoids is likely to result in simultaneous changes to both the concentration and composition of carotenoids allocated to signaling, because carotenoids can be converted to other forms after ingestion (Hill 1996; Wedekind et al. 1998) and there are often competing somatic demands for dietary carotenoids (Lozano 1994; Olson and Owens 1998; von Schantz et al. 1999; Vinkler and Albrecht 2010). Moreover, dietary manipulation of carotenoids is likely to change not only the expression of the carotenoid color, but also the animals’ behavior and the expression of other phenotypic cues that may also act to mediate the behavior of the receiver (Blount et al. 2003). These are not only difficult to avoid experimentally, but impossible to quantify exhaustively. Therefore, studies relying on dietary manipulation of carotenoids can never entirely rule out that the results are due to a factor correlated with carotenoid availability rather than the effect of carotenoid availability on the expression of the color itself (Blount et al. 2003; Pike, Blount, Bjerkeng, et al. 2007a). One potential methodological solution to this is to artificially manipulate the appearance of carotenoid-based colors using, for example, colored dyes or paints (Burley and Coopersmith 1987; Pryke and Andersson 2003; Kaldonski et al. 2009; Alonso-Alvarez et al. 2012), lighting deficient in the wavelengths needed to perceive variation in the signal (Long and Houde 1989; Milinski and Bakker 1990; Evans and Norris 1996; Rick and Bakker 2008), or computer-generated colors on animated animals (McKinnon and McPhail 1996; Bakker et al. 1999; Kunzler and Bakker 2001; Mazzi et al. 2003; Mazzi et al. 2004; Zbinden et al. 2004; Veen et al. 2013). However, while typically achieving the goal of manipulating color appearance, at least to human observers, in very few of these studies was any attempt made to generate signals that perceptually mimicked those they were emulating (although see Pryke and Andersson 2003; Kaldonski et al. 2009). The conclusions which follow from these manipulations must therefore be interpreted with caution, as the manipulated colors may be perceived and interpreted in unintended ways (Bennett et al. 1994). Moreover, no studies using artificial manipulation of carotenoid-based colors have attempted to replicate the range of natural variation in color on which female choice could operate, instead typically focusing on one or a small number of exemplar colors (Bakker et al. 1999; Pryke and Andersson 2003; Kaldonski et al. 2009). We therefore have a poor understanding of how females respond to biologically relevant variation in signal color, independently of other sources of information about signaler quality. In this article, I aim to overcome some of the limitations of previous studies by quantifying female choice for males that differ exclusively in the color of their carotenoid-based signal, and display biologically relevant signal colors representative of the multidimensional chromatic variation inherent in natural populations. Three-spined sticklebacks (Gasterosteus aculeatus) are a classic model in sexual selection studies looking at the evolution of carotenoid-based signals (Östlund-Nilsson 2007). During the breeding season, male sticklebacks develop a region of intense, carotenoid-based throat coloration that is used by females during mate-choice (Wootton 1984; Milinski and Bakker 1990), along with various other color-based, morphological and chemical traits (Kunzler and Bakker 2001; Candolin 2003; Flamarique et al. 2013). The carotenoid signal is composed predominantly of astaxanthin, lutein, and tunaxanthin fatty acyl esters, with tunaxanthin likely being metabolically derived from dietary astaxanthin (Wedekind et al. 1998). The precise color expressed is therefore a function of both the total concentration of carotenoids allocated to the signal and its composition, specifically regarding the relative proportion of the red-colored astaxanthin to the yellow/orange-colored tunaxanthin/lutein (Wedekind et al. 1998; Pike et al. 2011). Recent theoretical work (Pike et al. 2011) suggests that the 3-spined stickleback’s visual system is acutely sensitive to variation in the carotenoid content of this signal, with strong linear relationships between stickleback’s perception of color and both carotenoid concentration and composition, although this has not been demonstrated empirically. Indeed, the ability to distinguish between these 2 chromatic axes is predicted to be ecologically important for sticklebacks. Variation in signal “redness” (which is driven by variation in the relative concentration of astaxanthin) has been hypothesized to underpin species recognition (Rowe et al. 2004) and may allow males to maximize their contrast against the background (Baube et al. 1995; Boughman 2001) or exploit intrapopulation variation in female color preferences (Baube et al. 1995). In addition, variation in the concentration of carotenoids allocated to sexual signaling is known to be related to dietary availability, levels of somatic oxidative stress (Pike, Blount, Bjerkeng, et al. 2007a), and parasite infection (Milinski and Bakker 1990), and so females are likely to benefit by selecting potential mates on the basis of signal carotenoid content (Pike, Blount, Bjerkeng, et al. 2007a). Here, I utilized this relationship between the perceived color of a male’s carotenoid signal and its underlying carotenoid content (Pike et al. 2011) to create virtual, computer-animated males with nuptial coloration that perceptually matched that of live males with known carotenoid content. Computer animations of courting males are well-established in the study of stickleback mating preferences (McKinnon and McPhail 1996; Bakker et al. 1999; Kunzler and Bakker 2001; Mazzi et al. 2003; Mazzi et al. 2004; Zbinden et al. 2004; Veen et al. 2013), and have the benefit that the specific trait of interest (in this case perceived variation in allocation of carotenoids to a carotenoid-based sexual signal) can be manipulated independently of all others. However, in previous studies, colors were used that appeared appropriate to human observers, but made no attempt was made to ensure that colors appeared realistic to conspecific receivers. Since human and stickleback spectral sensitivities differ considerably, particularly towards the red end of the spectrum (Supplementary Figure S1), it is therefore difficult to interpret how these colors would have been perceived and interpreted by courting females. To overcome this, the carotenoid-based colors on the virtual males used here were designed to produce a metameric match for those measured spectrophotometrically, in terms of sticklebacks’ color perception (i.e., to the viewing female the carotenoid-based color on the monitor result in the same pattern of photoreceptor stimulation, and hence produce colors that appear perceptually identical, to those produced by the live male it was based on: Fleishman et al. 1998; Smith and Pokorny 1995; Westland and Ripamonti 2004). This allowed me to test 1) whether females are able to discriminate between males solely on the basis of variation in carotenoid allocation, and 2) which dimension of signal variation, carotenoid composition or carotenoid concentration, they use to inform their choice. METHODS Source of fish and holding conditions Adult 3-spined sticklebacks were caught with minnow traps and dip nets from the estuary of the Great Eau, a seminatural drainage network in Lincolnshire, UK, during early spring. The fish were subsequently maintained in several 25 L outdoor aquaria, and therefore exposed to natural variation in temperature and daylight, for at least 2 months prior to starting the experiment. They were fed to satiation daily on frozen bloodworm. Fifty percent water changes were performed once a week and the tanks were regularly cleaned. Stimulus construction Gravid females (n = 24) were selected as and when they became available, and given a single mate-choice trial in which they had to select between 2 courting virtual males that differed exclusively in the perceived color of their carotenoid-based throat signal. Specifically, females were shown virtual males with breeding coloration that mimicked, from the perspective of the observing female, spectrometrically-determined colors from live males for which detailed biochemical information on carotenoid composition (measured as the relative proportion of astaxanthin to tunaxanthin/lutein) and concentration (measured as μg total carotenoids per g skin) is available (see Pike et al. (2011) for full methodological details of the measurement of color and the determination of signal carotenoids in these males). Females were therefore exposed to a range of colors known to be produced by live stickleback males, and these were accurately represented on the monitor by producing colors that are a metameric match for those measured spectrophotometrically, in terms of sticklebacks’ color perception (see Figure 1 for the locations of these colors within stickleback chromaticity space, and the Supplementary Material for full details on the procedure for computing and generating the virtual colors). Note that sticklebacks are known to possess retinal photoreceptors sensitive to ultraviolet radiation, and the males’ carotenoid-based signal reflects within the UV region of the spectrum (Rick et al. 2004). Therefore, given the restricted gamut of the computer monitor (Supplementary Figure S1a) this information will necessarily be absent. However, recent work (Pike et al. 2011) suggests that sticklebacks can glean relevant information on the carotenoid content of this signal even in the complete absence of UV radiation, and indeed UV information is likely to be largely redundant for this task; in particular they found that a tetrachromatic model (which included a putative contribution from UV-sensitive cones) was no better at predicting carotenoid concentration or composition than a simpler, trichromatic model. This led them to conclude that only the stickleback’s 3 photoreceptors sensitive within the human-visible portion of the spectrum are needed to fully encode information on signal carotenoid content (Pike et al. 2011). There is therefore no reason to assume that an absence of UV information would preclude females from discriminating between males and extracting biologically meaningful information from the signal. Figure 1 View largeDownload slide The position of the 48 virtual males nuptial colors in stickleback chromaticity space (gray circles), in which the 3 apices of the triangle represent stimulation of the stickleback’s long (L), medium (M), and short (S) wavelength-sensitive cones in isolation (Kelber et al. 2003), after adaptation to the background radiance. Note that sticklebacks also have fourth, UV-sensitive cone class, although this is not included here as it did not contribute to color perception in this experiment. These colors have almost exactly the same spatial locations as a subset of those given for real males in Pike et al. (2011) and so were assumed to appear chromatically identical. The position of the virtual males’ iris color (white square) and flank color (white triangle) are also shown. The blue iris coloration was matched to that of an average of 23 measurements from breeding males from the study population, and the flank color was set to gray; fins were white and semitransparent and pupils black. Figure 1 View largeDownload slide The position of the 48 virtual males nuptial colors in stickleback chromaticity space (gray circles), in which the 3 apices of the triangle represent stimulation of the stickleback’s long (L), medium (M), and short (S) wavelength-sensitive cones in isolation (Kelber et al. 2003), after adaptation to the background radiance. Note that sticklebacks also have fourth, UV-sensitive cone class, although this is not included here as it did not contribute to color perception in this experiment. These colors have almost exactly the same spatial locations as a subset of those given for real males in Pike et al. (2011) and so were assumed to appear chromatically identical. The position of the virtual males’ iris color (white square) and flank color (white triangle) are also shown. The blue iris coloration was matched to that of an average of 23 measurements from breeding males from the study population, and the flank color was set to gray; fins were white and semitransparent and pupils black. In order to ensure that female mate-choice decisions were based solely on chromatic aspects of the signal, the “behavior” of the males as well as the size and perceived brightness of the carotenoid signal, and all other body colors (Figure 1) were standardized across males. Assuming that perceived brightness can be approximated by the sum of the output of all contributing cones classes (Endler and Mielke 2005; Pike et al. 2011), each carotenoid signal was designed to appear isoluminant to observing females by eliciting the same summed cone output; virtual males therefore differed exclusively in the color of their nuptial signal, with each indicative of a particular pattern of carotenoid allocation. Which specific pair of male colors a female saw was randomly assigned, and each male color was only used once. Preference tests Female preference between pairs of virtual males was tested using an established protocol in a dichotomous choice design (Bakker et al. 1999). A 33 × 18 × 19 cm glass aquarium tank, one end of which was divided into 2 equally sized “choice zones” by a 9 cm opaque partition protruding from the wall, was placed immediately in front of a calibrated 22-inch flat-screen CRT monitor (Iiyama VisonMaster 513), which displayed 2 static males at the “rear” of their virtual environment. This environment was uniformly gray and contained no landscape features. The rest of the monitor (i.e., everything apart from the 24 × 8 cm display window) was covered by black paper, as were the other 3 sides of the tank. The only available illumination was from the monitor. A single female was introduced into the tank and, after 5 min of acclimatization, was shown 2 simultaneous animation sequences, one presented in front of each choice zone, of a male performing a “zig-zag” courtship display in the direction of the viewing female and then returning directly to the starting position (see the Supplementary Material for full details on the construction and presentation of the animation sequences, and Supplementary Movie S1 for a representative animation). These videos were shown on a loop for 5 min, after which the screen went black and the trial ended. Throughout each trial, the female’s tank was video-recorded from above using a webcam linked to a computer, allowing her position to be tracked in real time using a custom-written Matlab (MathWorks, Natick, MA) function. The amount of time (s) any part of her body spent in the “choice zone” in front of each virtual male was automatically recorded and used to assess her mating preference, as the relative proportion of time spent with one male given the time she spent with either male (see below). Trait preferences recorded using this metric have been shown to correspond closely to those shown by females that are actually able to spawn with their preferred male (Cubillos and Guderley 2000). Statistical analysis Females’ response to variation in male coloration was analyzed using a randomization test based on fitting linear models using the “lm” function in R version 3.4.1. Each model included the arcsine-square root transformed proportion of time females spent with one, randomly selected, male of each pair as the response variable, and the difference in total carotenoid concentration and difference in the relative proportion of astaxanthin between this male and the alternative male (and their interaction) as continuous predictors. A positive relationship would therefore indicate that females expressed an overall preference for the male of a pair with the greatest perceived concentration of carotenoids, or the highest perceived proportion of astaxanthin, in their signal. Because the specific (random) choice of male may affect the strength of any observed relationships, I reran the analysis 1000 times, choosing a different random male from a pair at each iteration. F-ratios for the main effect and interaction terms were collated, and used to compute P values as the proportion of F ratios that were less than or equal to the appropriate critical value (which, for a model with 2 covariates and their interaction and a total sample size of 24, was 4.35 in each case). Mean F ratios and coefficients of variation (R2) for each main effect and interaction term, from across all randomizations of the data, are reported along with their associated bootstrap confidence intervals (Carpenter and Bithell 2000). The relationship between the difference in carotenoid concentration and the difference in carotenoid composition between pairs of males was tested using a randomization-based correlation, conducted as described above, but using the Pearson correlation coefficient (r) as the test statistic and comparing against a critical value of 0.404 (for 22 degrees of freedom). RESULTS Female preference was significantly predicted by the difference in total carotenoid concentration between the 2 males that she saw (randomization test: F1,20 = 8.60 [95% confidence interval: 7.94, 9.26], R2 = 0.29 [0.27, 0.31], P = 0.003; Figure 2a), but not by the difference in the relative proportion of astaxanthin (F1,20 = 0.07 [0.01, 0.13], R2 = 0.03 [0.02, 0.04], P = 0.997; Figure 2b) or their interaction (F1,20 = 0.56 [0.27, 0.85], R2 = 0.05 [0.03, 0.07], P = 0.995). The male of the pair with the greatest perceived concentration of carotenoids in his signal was therefore preferred, suggesting that female preference was determined solely by signal carotenoid concentration regardless of its composition. Figure 2 View largeDownload slide Representative female preference (proportion of time spent with one, randomly selected, male of each pair) as a function of the difference in (a) perceived total carotenoid concentration, and (b) perceived proportion of astaxanthin between the 2 males she saw. Preference scores > 0.5 indicate preference for the selected male; those < 0.5 indicate preference for the alternative male. Point size is proportional to the sum total amount of time females spent with either male (ranging from 45 to 254 s). In (a), the least-squares line of best fit is shown. Figure 2 View largeDownload slide Representative female preference (proportion of time spent with one, randomly selected, male of each pair) as a function of the difference in (a) perceived total carotenoid concentration, and (b) perceived proportion of astaxanthin between the 2 males she saw. Preference scores > 0.5 indicate preference for the selected male; those < 0.5 indicate preference for the alternative male. Point size is proportional to the sum total amount of time females spent with either male (ranging from 45 to 254 s). In (a), the least-squares line of best fit is shown. There was no significant correlation between the difference in carotenoid concentration and the difference in carotenoid composition between pairs of males (randomization test: r = 0.27 [0.25, 0.29], df = 22, P = 0.995; Supplementary Figure S2). DISCUSSION The results of this study show that female 3-spined sticklebacks, at least from this study population, show a preference for males allocating larger concentrations of carotenoids to the signal overall, rather than males varying the relative proportion of constituent carotenoids (Wedekind et al. 1998; Pike et al. 2011). They are therefore choosing males that, in human perceptual terms, are increasingly saturated rather than those that appear redder (Pike et al. 2011). By controlling for male behavior, while presenting biologically-relevant colors, we can be confident that the observed female preferences are mediated by color, rather than a correlated aspect of the male’s phenotype or behavior. This allows us to draw conclusions directly relating to the information content of the carotenoid-based color, and in particular to identify which dimension of the signal females are basing their choice on. The data from this study suggest that female sticklebacks’ long-postulated preference for “redness” in males’ sexual signals (Tinbergen 1951) may in fact be misleading, and that in actual fact any signal that is sufficiently rich in carotenoids, be it relatively high or low in astaxanthin, is likely to be viewed favorably by choosing females. However, this is not to say that carotenoid composition is not important in other contexts. Rowe et al. (2004) argued that variation along the red-orange axis (i.e., variation in carotenoid composition; Pike et al. (2011)) would be most useful in species recognition (i.e., using a redness threshold or range to categorize individuals as conspecifics or heterospecifics), and so males would be predicted to express colors within a given range of species norms. Moreover, variation in signal redness may allow males to increase their detectability by conspecifics, for example by maximizing contrast against the background or the efficacy of the signal’s transmission given the prevailing photic environment and optical properties of the water (Endler 1992; Baube et al. 1995; Boughman 2001; Morrongiello et al. 2010), or may be an adaptation to intra- or interpopulation variation in female preference for signal color. For example, some stickleback females have been reported to prefer orange or yellow over red males, perhaps because of condition-dependent visual constraints or differences in female “motivation” to choose (Baube et al. 1995; Bakker et al. 1999), and there is evidence from studies on guppies (Poecilia reticulata) that males may attempt to match a particular population-specific pigment ratio in their sexual signal to exploit female preference for color-based signals (Grether et al. 2005). However, while sticklebacks are known to adapt the chromatic aspects of their carotenoid-based signal over evolutionary and ecological time periods (Boughman 2001; Lindström et al. 2009), dynamic changes over shorter periods of time, for example in response to territorial intrusions from rival males (Candolin 1999; Kim and Velando 2014; Hiermes et al. 2016), tend to be in terms of signal brightness. These longer-term changes are almost certainly influenced, at least in part, by carotenoid allocation (e.g., the intensity of carotenoid signaling is often reduced over successive breeding rounds, possibly due to reabsorption or loss of carotenoids from the skin, and may be increased towards the end of the breeding season as a form of “terminal investment”; Lindström et al. (2009)). In contrast, any short-term changes that occur are most likely to result from physiological changes in the skin (e.g., the dispersion or aggregation of carotenoid-containing organelles; Grether et al. (2004)), rather than direct changes in the concentration or composition of the carotenoids themselves. One key difference between perceived variation in the composition and concentration of colors may be that they are exploited at different behaviorally-relevant spatial scales. For instance, signal redness/orangeness may be important in allowing the detection of a signal in the first place, probably from distances up to a few meters; only at closer ranges of a few centimeters (as in the current study) may females then pay attention to the concentration of carotenoids (Rowe et al. 2004). This is consistent with explanations for the involvement of carotenoid concentration in mate assessment and choice in this, and other species. In particular, variation between individuals in their ability to acquire, sequester or metabolize dietary carotenoids (Endler 1980; McGraw 2004; Hill and Johnson 2012), or in the extent to which these carotenoids are diverted away from signaling to other somatic functions, such as antioxidant defense and/or immune-support (Lozano 1994; Olson and Owens 1998; von Schantz et al. 1999; Vinkler and Albrecht 2010), have been repeatedly linked to their signaling function (Svensson and Wong 2011). Indeed, variation in the concentration of carotenoids allocated to sexual signaling in sticklebacks is known to be related to dietary availability, levels of somatic oxidative stress (Pike, Blount, Bjerkeng, et al. 2007a) and parasite infection (Milinski and Bakker 1990). Selecting potential mates on a signal linked directly to signaler quality may have significant fitness implications, especially in this species where males alone are responsible for all postlaying care of the eggs (Wootton 1984) and the quality of care they provide (and hence the number of offspring they raise to independence) is affected by dietary carotenoid availability (Pike, Blount, Lindstrom, et al. 2007b). As it is likely that the underlying concentration of carotenoids allocated to sexual signaling is an important determinant of female mating preferences across a wide range of taxa, this finding has important general implications for understanding male signaling strategies and the information content of carotenoid-based sexual signals. SUPPLEMENTARY MATERIAL Supplementary data are available at Behavioral Ecology online. FUNDING This research was supported by a Natural Environment Research Council fellowship (NE/F016514/2). The author thanks D. Masterton (The Design Centre, University College Falmouth) for scanning the fish model and for advice and discussion on the 3-dimensional modeling techniques. Data accessibility: Analyses reported in this article can be reproduced using the data provided by Pike (2017). REFERENCES Alonso-Alvarez C, Perez-Rodriguez L, Ferrero ME, de-Blas EG, Casas F, Mougeot F. 2012. Adjustment of female reproductive investment according to male carotenoid-based ornamentation in a gallinaceous bird. Behav Ecol Sociobiol . 66: 731– 742. Google Scholar CrossRef Search ADS   Bakker TCM, Kunzler R, Mazzi D. 1999. Sexual selection - Condition-related mate choice in sticklebacks. Nature . 401: 234– 234. Google Scholar CrossRef Search ADS   Baube CL, Rowland WJ, Fowler JB. 1995. The mechanisms of colour-based mate choice in female threespine sticklebacks: hue, contrast and configurational, cues. Behaviour . 132: 979– 996. Google Scholar CrossRef Search ADS   Bennett ATD, Cuthill IC, Norris KJ. 1994. Sexual selection and the mismeasure of color. Am Nat . 144: 848– 860. Google Scholar CrossRef Search ADS   Blount JD, Metcalfe NB, Birkhead TR, Surai PF. 2003. Carotenoid modulation of immune function and sexual attractiveness in zebra finches. Science . 300: 125– 127. Google Scholar CrossRef Search ADS PubMed  Boughman JW. 2001. Divergent sexual selection enhances reproductive isolation in sticklebacks. Nature . 411: 944– 948. Google Scholar CrossRef Search ADS PubMed  Britton G, Liaaen-Jensen S, Pfander H. 2004. Carotenoids. Handbook . Berlin: Birkhäuser Verlag. Google Scholar CrossRef Search ADS   Brush AH. 1990. Metabolism of carotenoid pigments in birds. FASEB J . 4: 2969– 2977. Google Scholar PubMed  Burley N, Coopersmith CB. 1987. Bill color preferences of zebra finches. Ethology . 76: 133– 151. Google Scholar CrossRef Search ADS   Candolin U. 1999. Male-male competition facilitates female choice in sticklebacks. P Roy Soc B-Biol Sci . 266: 785– 789. Google Scholar CrossRef Search ADS   Candolin U. 2003. The use of multiple cues in mate choice. Biol Rev Camb Philos Soc . 78: 575– 595. Google Scholar CrossRef Search ADS PubMed  Carpenter J, Bithell J. 2000. Bootstrap confidence intervals: when, which, what? A practical guide for medical statisticians. Stat Med . 19: 1141– 1164. Google Scholar CrossRef Search ADS PubMed  Cubillos ER, Guderley HE. 2000. Analysis of the factors related with mate choice and reproductive success in male three-spined sticklebacks. J Fish Biol . 56: 1201– 1216. Google Scholar CrossRef Search ADS   Endler JA. 1980. Natural selection on color patterns in Poecilia reticulata. Evolution . 34: 76– 91. Google Scholar CrossRef Search ADS PubMed  Endler JA. 1992. Signals, signal conditions, and the direction of evolution. Am Nat . 139: S125– S153. Google Scholar CrossRef Search ADS   Endler JA, Mielke PW. 2005. Comparing entire colour patterns as birds see them. Biol J Linn Soc . 86: 405– 431. Google Scholar CrossRef Search ADS   Evans MR, Norris K. 1996. The importance of carotenoids in signaling during aggressive interactions between male firemouth cichlids (Cichlasoma meeki). Behav Ecol . 7: 1– 6. Google Scholar CrossRef Search ADS   Flamarique IN, Bergstrom C, Cheng CL, Reimchen TE. 2013. Role of the iridescent eye in stickleback female mate choice. J Exp Biol . 216: 2806– 2812. Google Scholar CrossRef Search ADS PubMed  Fleishman LJ, McClintock WJ, D’Eath RB, Brainard DH, Endler JA. 1998. Colour perception and the use of video playback experiments in animal behaviour. Anim Behav. 56: 1035 – 1040. Goodwin TW. 1984. The biochemistry of the carotenoids, volume II: animals . London: Chapman & Hall. Google Scholar CrossRef Search ADS   Grether GF, Cummings ME, Hudon J. 2005. Countergradient variation in the sexual coloration of guppies (Poecilia reticulata): drosopterin synthesis balances carotenoid availability. Evolution . 59: 175– 188. Google Scholar CrossRef Search ADS PubMed  Grether GF, Kolluru GR, Nersissian K. 2004. Individual colour patches as multicomponent signals. Biol Rev Camb Philos Soc . 79: 583– 610. Google Scholar CrossRef Search ADS PubMed  Hiermes M, Rick IP, Mehlis M, Bakker TCM. 2016. The dynamics of color signals in male threespine sticklebacks Gasterosteus aculeatus. Curr Zool . 62: 23– 31. Google Scholar CrossRef Search ADS   Hill GE. 1996. Redness as a measure of the production cost of ornamental coloration. Ethol Ecol Evol . 8: 157– 175. Google Scholar CrossRef Search ADS   Hill GE, Johnson JD. 2012. The vitamin A-redox hypothesis: a biochemical basis for honest signaling via carotenoid pigmentation. Am Nat . 180: E127– E150. Google Scholar CrossRef Search ADS PubMed  Hill GE, McGraw KJ. 2006a. Bird coloration volume 1: mechanisms and measurements . Cambridge (MA): Harvard University Press. Hill GE, McGraw KJ. 2006b. Bird coloration volume 2: function and evolution . Cambridge (MA): Harvard University Press. Kaldonski N, Perrot-Minnot MJ, Dodet R, Martinaud G, Cézilly F. 2009. Carotenoid-based colour of acanthocephalan cystacanths plays no role in host manipulation. Proc Biol Sci . 276: 169– 176. Google Scholar CrossRef Search ADS PubMed  Kelber A, Vorobyev M, Osorio D. 2003. Animal colour vision–behavioural tests and physiological concepts. Biol Rev Camb Philos Soc . 78: 81– 118. Google Scholar CrossRef Search ADS PubMed  Kemp DJ, Grether GF. 2014. Integrating functional and evolutionary approaches to the study of color-based animal signals. In: Duncan J. Irschick, Mark Briffa, Jeffrey Podos, editors. Animal signaling and function: an integrative approach . Hoboken (NJ): Wiley-Blackwell. p. 111– 140. Google Scholar CrossRef Search ADS   Kim SY, Velando A. 2014. Stickleback males increase red coloration and courtship behaviours in the presence of a competitive rival. Ethology . 120: 502– 510. Google Scholar CrossRef Search ADS   Kunzler R, Bakker TCM. 2001. Female preferences for single and combined traits in computer animated stickleback males. Behav Ecol . 12: 681– 685. Google Scholar CrossRef Search ADS   Lindström J, Pike TW, Blount JD, Metcalfe NB. 2009. Optimization of resource allocation can explain the temporal dynamics and honesty of sexual signals. Am Nat . 174: 515– 525. Google Scholar CrossRef Search ADS PubMed  Long KD, Houde AE. 1989. Orange spots as a visual cue for female mate choice in the guppy (Poecilia reticulata). Ethology . 82: 316– 324. Google Scholar CrossRef Search ADS   Lozano GA. 1994. Carotenoids, parasites, and sexual selection. Oikos . 70: 309– 311. Google Scholar CrossRef Search ADS   Mazzi D, Kunzler R, Bakker TCM. 2003. Female preference for symmetry in computer-animated three-spined sticklebacks, Gasterosteus aculeatus. Behav Ecol Sociobiol . 54: 156– 161. Mazzi D, Künzler R, Largiadèr CR, Bakker TC. 2004. Inbreeding affects female preference for symmetry in computer-animated sticklebacks. Behav Genet . 34: 417– 424. Google Scholar CrossRef Search ADS PubMed  McGraw KJ. 2004. Colorful songbirds metabolize carotenoids at the integument. J Avian Biol . 35: 471– 476. Google Scholar CrossRef Search ADS   McGraw KJ, Adkins-Regan E, Parker RS. 2002. Anhydrolutein in the zebra finch: a new, metabolically-derived carotenoid in birds. Comp Biochem Phys B . 132: 811– 818. Google Scholar CrossRef Search ADS   McKinnon JS, McPhail JD. 1996. Male aggression and colour in divergent populations of the threespine stickleback: experiments with animations. Can J Zool . 74: 1727– 1733. Google Scholar CrossRef Search ADS   Milinski M, Bakker TCM. 1990. Female sticklebacks use male coloration in mate choice and hence avoid parasitized males. Nature . 344: 330– 333. Google Scholar CrossRef Search ADS   Morrongiello JR, Bond NR, Crook DA, Wong BB. 2010. Nuptial coloration varies with ambient light environment in a freshwater fish. J Evol Biol . 23: 2718– 2725. Google Scholar CrossRef Search ADS PubMed  Olson VA, Owens IP. 1998. Costly sexual signals: are carotenoids rare, risky or required? Trends Ecol Evol . 13: 510– 514. Google Scholar CrossRef Search ADS PubMed  Östlund-Nilsson S. 2007. Reproductive behaviour in the three-spined stickleback. In: Östlund-Nilsson S, Mayer I, Huntingford FA, editors. Biology of the three-spined stickleback . Boca Raton (FL): CRC Press. p. 157– 178. Pike TW. 2017. Data from: discrimination of signal carotenoid content using multidimensional chromatic information. Dryad Digital Repository . http://dx.doi.org/10.5061/dryad.rd7jt Pike TW, Bjerkeng B, Blount JD, Lindstrom J, Metcalfe NB. 2011. How integument colour reflects its carotenoid content: a stickleback’s perspective. Funct Ecol . 25: 297– 304. Google Scholar CrossRef Search ADS   Pike TW, Blount JD, Bjerkeng B, Lindström J, Metcalfe NB. 2007. Carotenoids, oxidative stress and female mating preference for longer lived males. Proc Biol Sci . 274: 1591– 1596. Google Scholar CrossRef Search ADS PubMed  Pike TW, Blount JD, Lindstrom J, Metcalfe NB. 2007. Dietary carotenoid availability influences a male’s ability to provide parental care. Behav Ecol . 18: 1100– 1105. Google Scholar CrossRef Search ADS   Pryke SR, Andersson S. 2003. Carotenoid-based status signalling in red-shouldered widowbirds (Euplectes axillaris): epaulet size and redness affect captive and territorial competition. Behav Ecol Sociobiol . 53: 393– 401. Rick IP, Bakker TC. 2008. Color signaling in conspicuous red sticklebacks: do ultraviolet signals surpass others? BMC Evol Biol . 8: 189. Google Scholar CrossRef Search ADS PubMed  Rick IP, Modarressie R, Bakker TCM. 2004. Male three-spined sticklebacks reflect in ultraviolet light. Behaviour . 141: 1531– 1541. Google Scholar CrossRef Search ADS   Romero-Diaz C, Richner H, Granado-Lorencio F, Tschirren B, Fitze PS. 2013. Independent sources of condition dependency and multiple pathways determine a composite trait: lessons from carotenoid-based plumage colouration. J Evol Biol . 26: 635– 646. Google Scholar CrossRef Search ADS PubMed  Rowe MP, Baube CL, Loew ER, Phillips JB. 2004. Optimal mechanisms for finding and selecting mates: how threespine stickleback (Gasterosteus aculeatus) should encode male throat colors. J Comp Physiol A . 190: 241– 256. Google Scholar CrossRef Search ADS   Smith VC, Pokorny J. 1995. Chromatic-discrimination axes, CRT phosphor spectra, and individual variation in color-vision. J Opt Soc Am A. 12: 27– 35. Stradi R, Hudon J, Celentano G, Pini E. 1998. Carotenoids in bird plumage: the complement of yellow and red pigments in true woodpeckers (Picinae). Comp Biochem Phys B . 120: 223– 230. Google Scholar CrossRef Search ADS   Svensson PA, Wong BBM. 2011. Carotenoid-based signals in behavioural ecology: a review. Behaviour . 148: 131– 189. Google Scholar CrossRef Search ADS   Tinbergen N. 1951. The study of instinct . London: Oxford University Press. Veen T, Ingley SJ, Cui RF, Simpson J, Asl MR, Zhang J, Butkowski T, Li W, Hash C, Johnson JBet al.   2013. anyFish: an open-source software to generate animated fish models for behavioural studies. Evol Ecol Res . 15: 361– 375. Vinkler M, Albrecht T. 2010. Carotenoid maintenance handicap and the physiology of carotenoid-based signalisation of health. Naturwissenschaften . 97: 19– 28. Google Scholar CrossRef Search ADS PubMed  von Schantz T, Bensch S, Grahn M, Hasselquist D, Wittzell H. 1999. Good genes, oxidative stress and condition-dependent sexual signals. Proc Biol Sci . 266: 1– 12. Google Scholar CrossRef Search ADS PubMed  Wedekind C, Meyer P, Frischknecht M, Niggli UA, Pfander H. 1998. Different carotenoids and potential information content of red coloration of male three-spined stickleback. J Chem Ecol . 24: 787– 801. Google Scholar CrossRef Search ADS   Westland S, Ripamonti C. 2004. Computational colour science using MATLAB. Chichester (UK): John Wiley. Wootton RJ. 1984. A functional biology of sticklebacks . Berkeley (CA): University of California Press. Google Scholar CrossRef Search ADS   Zbinden M, Largiader CR, Bakker TCM. 2004. Body size of virtual rivals affects ejaculate size in sticklebacks. Behav Ecol . 15: 137– 140. Google Scholar CrossRef Search ADS   © The Author(s) 2017. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Behavioral Ecology Oxford University Press

Discrimination of signal carotenoid content using multidimensional chromatic information

Loading next page...
 
/lp/ou_press/discrimination-of-signal-carotenoid-content-using-multidimensional-bS0uIRJwoc
Publisher
Oxford University Press
Copyright
© The Author(s) 2017. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1045-2249
eISSN
1465-7279
D.O.I.
10.1093/beheco/arx136
Publisher site
See Article on Publisher Site

Abstract

Abstract Red, orange, and yellow carotenoid-based ornaments, which are widely used as sexual signals in many birds, fish, and reptiles, are known to exhibit multidimensional chromatic variation as a result of both the concentration and relative proportions of different constituent carotenoids with differing spectral properties. This is thought to reflect intrinsic variation in signaler quality, making it a useful basis for female choice. However, whether females are able to discriminate relevant variation in carotenoid concentration and/or composition independently of each other, and of other phenotypic or behavior traits, and if so, how this mediates their choice, is poorly understood. Here, female 3-spined sticklebacks (Gasterosteus aculeatus) were presented with computer-animated courting males that varied exclusively in the appearance of their carotenoid-based coloration; specifically, each male’s signal provided a perceptual match for carotenoid coloration expressed by live males with known underlying carotenoid content, thereby providing a biologically-relevant signal while precluding confounding traits influencing female choice. Females were able to discriminate between prospective mates solely on the basis of perceived variation in the allocation of carotenoids to males’ sexual signals, and exhibited a strong preference for males with coloration indicative of higher concentrations of carotenoids in their signal, rather than in response to perceived variation in the relative proportion of constituent carotenoids. This has important implications for our understanding of male signaling strategies and the information content of carotenoid-based sexual signals. INTRODUCTION Color-based signals are one of the most widely used models in the study of sexual selection, and many studies have shown how individual variation in signal color can affect reproductive success through mate choice and intrasexual competition (Hill and McGraw 2006b; Kemp and Grether 2014). In these studies, color patches are typically considered either as single traits or as individual components of complex (multicomponent or multimodal) signals, and it is often overlooked that they themselves may be perceived as multidimensional signals, with different signal dimensions limited by specific constraints and subject to specific selection pressures (Candolin 2003; Grether et al. 2004). One good example of this is carotenoid-based coloration, which is responsible for much of the yellow, orange, and red regions of coloration present on reptiles, fish, and birds, and is known to exhibit multidimensional chromatic variation (Grether et al. 2004; Rowe et al. 2004; Pike et al. 2011; Romero-Diaz et al. 2013) as a result of variation in both the concentration and identity of the constituent carotenoids, which can differ in their spectral properties (Britton et al. 2004). Pigmented skin and feathers, for example, typically contain several different carotenoid pigments, and variation in the relative concentration of each can markedly affect the color produced (Stradi et al. 1998). Because vertebrates are incapable of producing carotenoid pigments de novo, dietary carotenoids are often used directly, and unchanged, in signaling (Olson and Owens 1998). However, in most species, spectrally distinct forms can be derived from the metabolic transformation of ingested carotenoids (Goodwin 1984; Brush 1990; Stradi et al. 1998; Wedekind et al. 1998; McGraw et al. 2002; Pike et al. 2011), potentially allowing animals control over the appearance of carotenoid-based colors and providing a mechanistic basis for the substantial variation observed both between and within species (Hill and McGraw 2006a). Most studies of carotenoid signaling implicitly assume that a receiver’s decisions are mediated ultimately by this underlying variation in carotenoid allocation, inferred through variation in the color produced. However, while we have a good understanding of the observed variation in carotenoid-based signal expression (Hill and McGraw 2006b), and of the biochemical basis for this variation (Britton et al. 2004), we have a comparatively poor understanding of whether the underlying patterns of carotenoid allocation are actually detectable by receivers and, if they are, how these perceived differences influence behavior. One of the reasons for this ambiguity is due to the inherent difficulty in manipulating different dimensions of the same signal independently in live animals without inadvertently affecting the expression of other cues. For example, dietary manipulation of carotenoids is likely to result in simultaneous changes to both the concentration and composition of carotenoids allocated to signaling, because carotenoids can be converted to other forms after ingestion (Hill 1996; Wedekind et al. 1998) and there are often competing somatic demands for dietary carotenoids (Lozano 1994; Olson and Owens 1998; von Schantz et al. 1999; Vinkler and Albrecht 2010). Moreover, dietary manipulation of carotenoids is likely to change not only the expression of the carotenoid color, but also the animals’ behavior and the expression of other phenotypic cues that may also act to mediate the behavior of the receiver (Blount et al. 2003). These are not only difficult to avoid experimentally, but impossible to quantify exhaustively. Therefore, studies relying on dietary manipulation of carotenoids can never entirely rule out that the results are due to a factor correlated with carotenoid availability rather than the effect of carotenoid availability on the expression of the color itself (Blount et al. 2003; Pike, Blount, Bjerkeng, et al. 2007a). One potential methodological solution to this is to artificially manipulate the appearance of carotenoid-based colors using, for example, colored dyes or paints (Burley and Coopersmith 1987; Pryke and Andersson 2003; Kaldonski et al. 2009; Alonso-Alvarez et al. 2012), lighting deficient in the wavelengths needed to perceive variation in the signal (Long and Houde 1989; Milinski and Bakker 1990; Evans and Norris 1996; Rick and Bakker 2008), or computer-generated colors on animated animals (McKinnon and McPhail 1996; Bakker et al. 1999; Kunzler and Bakker 2001; Mazzi et al. 2003; Mazzi et al. 2004; Zbinden et al. 2004; Veen et al. 2013). However, while typically achieving the goal of manipulating color appearance, at least to human observers, in very few of these studies was any attempt made to generate signals that perceptually mimicked those they were emulating (although see Pryke and Andersson 2003; Kaldonski et al. 2009). The conclusions which follow from these manipulations must therefore be interpreted with caution, as the manipulated colors may be perceived and interpreted in unintended ways (Bennett et al. 1994). Moreover, no studies using artificial manipulation of carotenoid-based colors have attempted to replicate the range of natural variation in color on which female choice could operate, instead typically focusing on one or a small number of exemplar colors (Bakker et al. 1999; Pryke and Andersson 2003; Kaldonski et al. 2009). We therefore have a poor understanding of how females respond to biologically relevant variation in signal color, independently of other sources of information about signaler quality. In this article, I aim to overcome some of the limitations of previous studies by quantifying female choice for males that differ exclusively in the color of their carotenoid-based signal, and display biologically relevant signal colors representative of the multidimensional chromatic variation inherent in natural populations. Three-spined sticklebacks (Gasterosteus aculeatus) are a classic model in sexual selection studies looking at the evolution of carotenoid-based signals (Östlund-Nilsson 2007). During the breeding season, male sticklebacks develop a region of intense, carotenoid-based throat coloration that is used by females during mate-choice (Wootton 1984; Milinski and Bakker 1990), along with various other color-based, morphological and chemical traits (Kunzler and Bakker 2001; Candolin 2003; Flamarique et al. 2013). The carotenoid signal is composed predominantly of astaxanthin, lutein, and tunaxanthin fatty acyl esters, with tunaxanthin likely being metabolically derived from dietary astaxanthin (Wedekind et al. 1998). The precise color expressed is therefore a function of both the total concentration of carotenoids allocated to the signal and its composition, specifically regarding the relative proportion of the red-colored astaxanthin to the yellow/orange-colored tunaxanthin/lutein (Wedekind et al. 1998; Pike et al. 2011). Recent theoretical work (Pike et al. 2011) suggests that the 3-spined stickleback’s visual system is acutely sensitive to variation in the carotenoid content of this signal, with strong linear relationships between stickleback’s perception of color and both carotenoid concentration and composition, although this has not been demonstrated empirically. Indeed, the ability to distinguish between these 2 chromatic axes is predicted to be ecologically important for sticklebacks. Variation in signal “redness” (which is driven by variation in the relative concentration of astaxanthin) has been hypothesized to underpin species recognition (Rowe et al. 2004) and may allow males to maximize their contrast against the background (Baube et al. 1995; Boughman 2001) or exploit intrapopulation variation in female color preferences (Baube et al. 1995). In addition, variation in the concentration of carotenoids allocated to sexual signaling is known to be related to dietary availability, levels of somatic oxidative stress (Pike, Blount, Bjerkeng, et al. 2007a), and parasite infection (Milinski and Bakker 1990), and so females are likely to benefit by selecting potential mates on the basis of signal carotenoid content (Pike, Blount, Bjerkeng, et al. 2007a). Here, I utilized this relationship between the perceived color of a male’s carotenoid signal and its underlying carotenoid content (Pike et al. 2011) to create virtual, computer-animated males with nuptial coloration that perceptually matched that of live males with known carotenoid content. Computer animations of courting males are well-established in the study of stickleback mating preferences (McKinnon and McPhail 1996; Bakker et al. 1999; Kunzler and Bakker 2001; Mazzi et al. 2003; Mazzi et al. 2004; Zbinden et al. 2004; Veen et al. 2013), and have the benefit that the specific trait of interest (in this case perceived variation in allocation of carotenoids to a carotenoid-based sexual signal) can be manipulated independently of all others. However, in previous studies, colors were used that appeared appropriate to human observers, but made no attempt was made to ensure that colors appeared realistic to conspecific receivers. Since human and stickleback spectral sensitivities differ considerably, particularly towards the red end of the spectrum (Supplementary Figure S1), it is therefore difficult to interpret how these colors would have been perceived and interpreted by courting females. To overcome this, the carotenoid-based colors on the virtual males used here were designed to produce a metameric match for those measured spectrophotometrically, in terms of sticklebacks’ color perception (i.e., to the viewing female the carotenoid-based color on the monitor result in the same pattern of photoreceptor stimulation, and hence produce colors that appear perceptually identical, to those produced by the live male it was based on: Fleishman et al. 1998; Smith and Pokorny 1995; Westland and Ripamonti 2004). This allowed me to test 1) whether females are able to discriminate between males solely on the basis of variation in carotenoid allocation, and 2) which dimension of signal variation, carotenoid composition or carotenoid concentration, they use to inform their choice. METHODS Source of fish and holding conditions Adult 3-spined sticklebacks were caught with minnow traps and dip nets from the estuary of the Great Eau, a seminatural drainage network in Lincolnshire, UK, during early spring. The fish were subsequently maintained in several 25 L outdoor aquaria, and therefore exposed to natural variation in temperature and daylight, for at least 2 months prior to starting the experiment. They were fed to satiation daily on frozen bloodworm. Fifty percent water changes were performed once a week and the tanks were regularly cleaned. Stimulus construction Gravid females (n = 24) were selected as and when they became available, and given a single mate-choice trial in which they had to select between 2 courting virtual males that differed exclusively in the perceived color of their carotenoid-based throat signal. Specifically, females were shown virtual males with breeding coloration that mimicked, from the perspective of the observing female, spectrometrically-determined colors from live males for which detailed biochemical information on carotenoid composition (measured as the relative proportion of astaxanthin to tunaxanthin/lutein) and concentration (measured as μg total carotenoids per g skin) is available (see Pike et al. (2011) for full methodological details of the measurement of color and the determination of signal carotenoids in these males). Females were therefore exposed to a range of colors known to be produced by live stickleback males, and these were accurately represented on the monitor by producing colors that are a metameric match for those measured spectrophotometrically, in terms of sticklebacks’ color perception (see Figure 1 for the locations of these colors within stickleback chromaticity space, and the Supplementary Material for full details on the procedure for computing and generating the virtual colors). Note that sticklebacks are known to possess retinal photoreceptors sensitive to ultraviolet radiation, and the males’ carotenoid-based signal reflects within the UV region of the spectrum (Rick et al. 2004). Therefore, given the restricted gamut of the computer monitor (Supplementary Figure S1a) this information will necessarily be absent. However, recent work (Pike et al. 2011) suggests that sticklebacks can glean relevant information on the carotenoid content of this signal even in the complete absence of UV radiation, and indeed UV information is likely to be largely redundant for this task; in particular they found that a tetrachromatic model (which included a putative contribution from UV-sensitive cones) was no better at predicting carotenoid concentration or composition than a simpler, trichromatic model. This led them to conclude that only the stickleback’s 3 photoreceptors sensitive within the human-visible portion of the spectrum are needed to fully encode information on signal carotenoid content (Pike et al. 2011). There is therefore no reason to assume that an absence of UV information would preclude females from discriminating between males and extracting biologically meaningful information from the signal. Figure 1 View largeDownload slide The position of the 48 virtual males nuptial colors in stickleback chromaticity space (gray circles), in which the 3 apices of the triangle represent stimulation of the stickleback’s long (L), medium (M), and short (S) wavelength-sensitive cones in isolation (Kelber et al. 2003), after adaptation to the background radiance. Note that sticklebacks also have fourth, UV-sensitive cone class, although this is not included here as it did not contribute to color perception in this experiment. These colors have almost exactly the same spatial locations as a subset of those given for real males in Pike et al. (2011) and so were assumed to appear chromatically identical. The position of the virtual males’ iris color (white square) and flank color (white triangle) are also shown. The blue iris coloration was matched to that of an average of 23 measurements from breeding males from the study population, and the flank color was set to gray; fins were white and semitransparent and pupils black. Figure 1 View largeDownload slide The position of the 48 virtual males nuptial colors in stickleback chromaticity space (gray circles), in which the 3 apices of the triangle represent stimulation of the stickleback’s long (L), medium (M), and short (S) wavelength-sensitive cones in isolation (Kelber et al. 2003), after adaptation to the background radiance. Note that sticklebacks also have fourth, UV-sensitive cone class, although this is not included here as it did not contribute to color perception in this experiment. These colors have almost exactly the same spatial locations as a subset of those given for real males in Pike et al. (2011) and so were assumed to appear chromatically identical. The position of the virtual males’ iris color (white square) and flank color (white triangle) are also shown. The blue iris coloration was matched to that of an average of 23 measurements from breeding males from the study population, and the flank color was set to gray; fins were white and semitransparent and pupils black. In order to ensure that female mate-choice decisions were based solely on chromatic aspects of the signal, the “behavior” of the males as well as the size and perceived brightness of the carotenoid signal, and all other body colors (Figure 1) were standardized across males. Assuming that perceived brightness can be approximated by the sum of the output of all contributing cones classes (Endler and Mielke 2005; Pike et al. 2011), each carotenoid signal was designed to appear isoluminant to observing females by eliciting the same summed cone output; virtual males therefore differed exclusively in the color of their nuptial signal, with each indicative of a particular pattern of carotenoid allocation. Which specific pair of male colors a female saw was randomly assigned, and each male color was only used once. Preference tests Female preference between pairs of virtual males was tested using an established protocol in a dichotomous choice design (Bakker et al. 1999). A 33 × 18 × 19 cm glass aquarium tank, one end of which was divided into 2 equally sized “choice zones” by a 9 cm opaque partition protruding from the wall, was placed immediately in front of a calibrated 22-inch flat-screen CRT monitor (Iiyama VisonMaster 513), which displayed 2 static males at the “rear” of their virtual environment. This environment was uniformly gray and contained no landscape features. The rest of the monitor (i.e., everything apart from the 24 × 8 cm display window) was covered by black paper, as were the other 3 sides of the tank. The only available illumination was from the monitor. A single female was introduced into the tank and, after 5 min of acclimatization, was shown 2 simultaneous animation sequences, one presented in front of each choice zone, of a male performing a “zig-zag” courtship display in the direction of the viewing female and then returning directly to the starting position (see the Supplementary Material for full details on the construction and presentation of the animation sequences, and Supplementary Movie S1 for a representative animation). These videos were shown on a loop for 5 min, after which the screen went black and the trial ended. Throughout each trial, the female’s tank was video-recorded from above using a webcam linked to a computer, allowing her position to be tracked in real time using a custom-written Matlab (MathWorks, Natick, MA) function. The amount of time (s) any part of her body spent in the “choice zone” in front of each virtual male was automatically recorded and used to assess her mating preference, as the relative proportion of time spent with one male given the time she spent with either male (see below). Trait preferences recorded using this metric have been shown to correspond closely to those shown by females that are actually able to spawn with their preferred male (Cubillos and Guderley 2000). Statistical analysis Females’ response to variation in male coloration was analyzed using a randomization test based on fitting linear models using the “lm” function in R version 3.4.1. Each model included the arcsine-square root transformed proportion of time females spent with one, randomly selected, male of each pair as the response variable, and the difference in total carotenoid concentration and difference in the relative proportion of astaxanthin between this male and the alternative male (and their interaction) as continuous predictors. A positive relationship would therefore indicate that females expressed an overall preference for the male of a pair with the greatest perceived concentration of carotenoids, or the highest perceived proportion of astaxanthin, in their signal. Because the specific (random) choice of male may affect the strength of any observed relationships, I reran the analysis 1000 times, choosing a different random male from a pair at each iteration. F-ratios for the main effect and interaction terms were collated, and used to compute P values as the proportion of F ratios that were less than or equal to the appropriate critical value (which, for a model with 2 covariates and their interaction and a total sample size of 24, was 4.35 in each case). Mean F ratios and coefficients of variation (R2) for each main effect and interaction term, from across all randomizations of the data, are reported along with their associated bootstrap confidence intervals (Carpenter and Bithell 2000). The relationship between the difference in carotenoid concentration and the difference in carotenoid composition between pairs of males was tested using a randomization-based correlation, conducted as described above, but using the Pearson correlation coefficient (r) as the test statistic and comparing against a critical value of 0.404 (for 22 degrees of freedom). RESULTS Female preference was significantly predicted by the difference in total carotenoid concentration between the 2 males that she saw (randomization test: F1,20 = 8.60 [95% confidence interval: 7.94, 9.26], R2 = 0.29 [0.27, 0.31], P = 0.003; Figure 2a), but not by the difference in the relative proportion of astaxanthin (F1,20 = 0.07 [0.01, 0.13], R2 = 0.03 [0.02, 0.04], P = 0.997; Figure 2b) or their interaction (F1,20 = 0.56 [0.27, 0.85], R2 = 0.05 [0.03, 0.07], P = 0.995). The male of the pair with the greatest perceived concentration of carotenoids in his signal was therefore preferred, suggesting that female preference was determined solely by signal carotenoid concentration regardless of its composition. Figure 2 View largeDownload slide Representative female preference (proportion of time spent with one, randomly selected, male of each pair) as a function of the difference in (a) perceived total carotenoid concentration, and (b) perceived proportion of astaxanthin between the 2 males she saw. Preference scores > 0.5 indicate preference for the selected male; those < 0.5 indicate preference for the alternative male. Point size is proportional to the sum total amount of time females spent with either male (ranging from 45 to 254 s). In (a), the least-squares line of best fit is shown. Figure 2 View largeDownload slide Representative female preference (proportion of time spent with one, randomly selected, male of each pair) as a function of the difference in (a) perceived total carotenoid concentration, and (b) perceived proportion of astaxanthin between the 2 males she saw. Preference scores > 0.5 indicate preference for the selected male; those < 0.5 indicate preference for the alternative male. Point size is proportional to the sum total amount of time females spent with either male (ranging from 45 to 254 s). In (a), the least-squares line of best fit is shown. There was no significant correlation between the difference in carotenoid concentration and the difference in carotenoid composition between pairs of males (randomization test: r = 0.27 [0.25, 0.29], df = 22, P = 0.995; Supplementary Figure S2). DISCUSSION The results of this study show that female 3-spined sticklebacks, at least from this study population, show a preference for males allocating larger concentrations of carotenoids to the signal overall, rather than males varying the relative proportion of constituent carotenoids (Wedekind et al. 1998; Pike et al. 2011). They are therefore choosing males that, in human perceptual terms, are increasingly saturated rather than those that appear redder (Pike et al. 2011). By controlling for male behavior, while presenting biologically-relevant colors, we can be confident that the observed female preferences are mediated by color, rather than a correlated aspect of the male’s phenotype or behavior. This allows us to draw conclusions directly relating to the information content of the carotenoid-based color, and in particular to identify which dimension of the signal females are basing their choice on. The data from this study suggest that female sticklebacks’ long-postulated preference for “redness” in males’ sexual signals (Tinbergen 1951) may in fact be misleading, and that in actual fact any signal that is sufficiently rich in carotenoids, be it relatively high or low in astaxanthin, is likely to be viewed favorably by choosing females. However, this is not to say that carotenoid composition is not important in other contexts. Rowe et al. (2004) argued that variation along the red-orange axis (i.e., variation in carotenoid composition; Pike et al. (2011)) would be most useful in species recognition (i.e., using a redness threshold or range to categorize individuals as conspecifics or heterospecifics), and so males would be predicted to express colors within a given range of species norms. Moreover, variation in signal redness may allow males to increase their detectability by conspecifics, for example by maximizing contrast against the background or the efficacy of the signal’s transmission given the prevailing photic environment and optical properties of the water (Endler 1992; Baube et al. 1995; Boughman 2001; Morrongiello et al. 2010), or may be an adaptation to intra- or interpopulation variation in female preference for signal color. For example, some stickleback females have been reported to prefer orange or yellow over red males, perhaps because of condition-dependent visual constraints or differences in female “motivation” to choose (Baube et al. 1995; Bakker et al. 1999), and there is evidence from studies on guppies (Poecilia reticulata) that males may attempt to match a particular population-specific pigment ratio in their sexual signal to exploit female preference for color-based signals (Grether et al. 2005). However, while sticklebacks are known to adapt the chromatic aspects of their carotenoid-based signal over evolutionary and ecological time periods (Boughman 2001; Lindström et al. 2009), dynamic changes over shorter periods of time, for example in response to territorial intrusions from rival males (Candolin 1999; Kim and Velando 2014; Hiermes et al. 2016), tend to be in terms of signal brightness. These longer-term changes are almost certainly influenced, at least in part, by carotenoid allocation (e.g., the intensity of carotenoid signaling is often reduced over successive breeding rounds, possibly due to reabsorption or loss of carotenoids from the skin, and may be increased towards the end of the breeding season as a form of “terminal investment”; Lindström et al. (2009)). In contrast, any short-term changes that occur are most likely to result from physiological changes in the skin (e.g., the dispersion or aggregation of carotenoid-containing organelles; Grether et al. (2004)), rather than direct changes in the concentration or composition of the carotenoids themselves. One key difference between perceived variation in the composition and concentration of colors may be that they are exploited at different behaviorally-relevant spatial scales. For instance, signal redness/orangeness may be important in allowing the detection of a signal in the first place, probably from distances up to a few meters; only at closer ranges of a few centimeters (as in the current study) may females then pay attention to the concentration of carotenoids (Rowe et al. 2004). This is consistent with explanations for the involvement of carotenoid concentration in mate assessment and choice in this, and other species. In particular, variation between individuals in their ability to acquire, sequester or metabolize dietary carotenoids (Endler 1980; McGraw 2004; Hill and Johnson 2012), or in the extent to which these carotenoids are diverted away from signaling to other somatic functions, such as antioxidant defense and/or immune-support (Lozano 1994; Olson and Owens 1998; von Schantz et al. 1999; Vinkler and Albrecht 2010), have been repeatedly linked to their signaling function (Svensson and Wong 2011). Indeed, variation in the concentration of carotenoids allocated to sexual signaling in sticklebacks is known to be related to dietary availability, levels of somatic oxidative stress (Pike, Blount, Bjerkeng, et al. 2007a) and parasite infection (Milinski and Bakker 1990). Selecting potential mates on a signal linked directly to signaler quality may have significant fitness implications, especially in this species where males alone are responsible for all postlaying care of the eggs (Wootton 1984) and the quality of care they provide (and hence the number of offspring they raise to independence) is affected by dietary carotenoid availability (Pike, Blount, Lindstrom, et al. 2007b). As it is likely that the underlying concentration of carotenoids allocated to sexual signaling is an important determinant of female mating preferences across a wide range of taxa, this finding has important general implications for understanding male signaling strategies and the information content of carotenoid-based sexual signals. SUPPLEMENTARY MATERIAL Supplementary data are available at Behavioral Ecology online. FUNDING This research was supported by a Natural Environment Research Council fellowship (NE/F016514/2). The author thanks D. Masterton (The Design Centre, University College Falmouth) for scanning the fish model and for advice and discussion on the 3-dimensional modeling techniques. Data accessibility: Analyses reported in this article can be reproduced using the data provided by Pike (2017). REFERENCES Alonso-Alvarez C, Perez-Rodriguez L, Ferrero ME, de-Blas EG, Casas F, Mougeot F. 2012. Adjustment of female reproductive investment according to male carotenoid-based ornamentation in a gallinaceous bird. Behav Ecol Sociobiol . 66: 731– 742. Google Scholar CrossRef Search ADS   Bakker TCM, Kunzler R, Mazzi D. 1999. Sexual selection - Condition-related mate choice in sticklebacks. Nature . 401: 234– 234. Google Scholar CrossRef Search ADS   Baube CL, Rowland WJ, Fowler JB. 1995. The mechanisms of colour-based mate choice in female threespine sticklebacks: hue, contrast and configurational, cues. Behaviour . 132: 979– 996. Google Scholar CrossRef Search ADS   Bennett ATD, Cuthill IC, Norris KJ. 1994. Sexual selection and the mismeasure of color. Am Nat . 144: 848– 860. Google Scholar CrossRef Search ADS   Blount JD, Metcalfe NB, Birkhead TR, Surai PF. 2003. Carotenoid modulation of immune function and sexual attractiveness in zebra finches. Science . 300: 125– 127. Google Scholar CrossRef Search ADS PubMed  Boughman JW. 2001. Divergent sexual selection enhances reproductive isolation in sticklebacks. Nature . 411: 944– 948. Google Scholar CrossRef Search ADS PubMed  Britton G, Liaaen-Jensen S, Pfander H. 2004. Carotenoids. Handbook . Berlin: Birkhäuser Verlag. Google Scholar CrossRef Search ADS   Brush AH. 1990. Metabolism of carotenoid pigments in birds. FASEB J . 4: 2969– 2977. Google Scholar PubMed  Burley N, Coopersmith CB. 1987. Bill color preferences of zebra finches. Ethology . 76: 133– 151. Google Scholar CrossRef Search ADS   Candolin U. 1999. Male-male competition facilitates female choice in sticklebacks. P Roy Soc B-Biol Sci . 266: 785– 789. Google Scholar CrossRef Search ADS   Candolin U. 2003. The use of multiple cues in mate choice. Biol Rev Camb Philos Soc . 78: 575– 595. Google Scholar CrossRef Search ADS PubMed  Carpenter J, Bithell J. 2000. Bootstrap confidence intervals: when, which, what? A practical guide for medical statisticians. Stat Med . 19: 1141– 1164. Google Scholar CrossRef Search ADS PubMed  Cubillos ER, Guderley HE. 2000. Analysis of the factors related with mate choice and reproductive success in male three-spined sticklebacks. J Fish Biol . 56: 1201– 1216. Google Scholar CrossRef Search ADS   Endler JA. 1980. Natural selection on color patterns in Poecilia reticulata. Evolution . 34: 76– 91. Google Scholar CrossRef Search ADS PubMed  Endler JA. 1992. Signals, signal conditions, and the direction of evolution. Am Nat . 139: S125– S153. Google Scholar CrossRef Search ADS   Endler JA, Mielke PW. 2005. Comparing entire colour patterns as birds see them. Biol J Linn Soc . 86: 405– 431. Google Scholar CrossRef Search ADS   Evans MR, Norris K. 1996. The importance of carotenoids in signaling during aggressive interactions between male firemouth cichlids (Cichlasoma meeki). Behav Ecol . 7: 1– 6. Google Scholar CrossRef Search ADS   Flamarique IN, Bergstrom C, Cheng CL, Reimchen TE. 2013. Role of the iridescent eye in stickleback female mate choice. J Exp Biol . 216: 2806– 2812. Google Scholar CrossRef Search ADS PubMed  Fleishman LJ, McClintock WJ, D’Eath RB, Brainard DH, Endler JA. 1998. Colour perception and the use of video playback experiments in animal behaviour. Anim Behav. 56: 1035 – 1040. Goodwin TW. 1984. The biochemistry of the carotenoids, volume II: animals . London: Chapman & Hall. Google Scholar CrossRef Search ADS   Grether GF, Cummings ME, Hudon J. 2005. Countergradient variation in the sexual coloration of guppies (Poecilia reticulata): drosopterin synthesis balances carotenoid availability. Evolution . 59: 175– 188. Google Scholar CrossRef Search ADS PubMed  Grether GF, Kolluru GR, Nersissian K. 2004. Individual colour patches as multicomponent signals. Biol Rev Camb Philos Soc . 79: 583– 610. Google Scholar CrossRef Search ADS PubMed  Hiermes M, Rick IP, Mehlis M, Bakker TCM. 2016. The dynamics of color signals in male threespine sticklebacks Gasterosteus aculeatus. Curr Zool . 62: 23– 31. Google Scholar CrossRef Search ADS   Hill GE. 1996. Redness as a measure of the production cost of ornamental coloration. Ethol Ecol Evol . 8: 157– 175. Google Scholar CrossRef Search ADS   Hill GE, Johnson JD. 2012. The vitamin A-redox hypothesis: a biochemical basis for honest signaling via carotenoid pigmentation. Am Nat . 180: E127– E150. Google Scholar CrossRef Search ADS PubMed  Hill GE, McGraw KJ. 2006a. Bird coloration volume 1: mechanisms and measurements . Cambridge (MA): Harvard University Press. Hill GE, McGraw KJ. 2006b. Bird coloration volume 2: function and evolution . Cambridge (MA): Harvard University Press. Kaldonski N, Perrot-Minnot MJ, Dodet R, Martinaud G, Cézilly F. 2009. Carotenoid-based colour of acanthocephalan cystacanths plays no role in host manipulation. Proc Biol Sci . 276: 169– 176. Google Scholar CrossRef Search ADS PubMed  Kelber A, Vorobyev M, Osorio D. 2003. Animal colour vision–behavioural tests and physiological concepts. Biol Rev Camb Philos Soc . 78: 81– 118. Google Scholar CrossRef Search ADS PubMed  Kemp DJ, Grether GF. 2014. Integrating functional and evolutionary approaches to the study of color-based animal signals. In: Duncan J. Irschick, Mark Briffa, Jeffrey Podos, editors. Animal signaling and function: an integrative approach . Hoboken (NJ): Wiley-Blackwell. p. 111– 140. Google Scholar CrossRef Search ADS   Kim SY, Velando A. 2014. Stickleback males increase red coloration and courtship behaviours in the presence of a competitive rival. Ethology . 120: 502– 510. Google Scholar CrossRef Search ADS   Kunzler R, Bakker TCM. 2001. Female preferences for single and combined traits in computer animated stickleback males. Behav Ecol . 12: 681– 685. Google Scholar CrossRef Search ADS   Lindström J, Pike TW, Blount JD, Metcalfe NB. 2009. Optimization of resource allocation can explain the temporal dynamics and honesty of sexual signals. Am Nat . 174: 515– 525. Google Scholar CrossRef Search ADS PubMed  Long KD, Houde AE. 1989. Orange spots as a visual cue for female mate choice in the guppy (Poecilia reticulata). Ethology . 82: 316– 324. Google Scholar CrossRef Search ADS   Lozano GA. 1994. Carotenoids, parasites, and sexual selection. Oikos . 70: 309– 311. Google Scholar CrossRef Search ADS   Mazzi D, Kunzler R, Bakker TCM. 2003. Female preference for symmetry in computer-animated three-spined sticklebacks, Gasterosteus aculeatus. Behav Ecol Sociobiol . 54: 156– 161. Mazzi D, Künzler R, Largiadèr CR, Bakker TC. 2004. Inbreeding affects female preference for symmetry in computer-animated sticklebacks. Behav Genet . 34: 417– 424. Google Scholar CrossRef Search ADS PubMed  McGraw KJ. 2004. Colorful songbirds metabolize carotenoids at the integument. J Avian Biol . 35: 471– 476. Google Scholar CrossRef Search ADS   McGraw KJ, Adkins-Regan E, Parker RS. 2002. Anhydrolutein in the zebra finch: a new, metabolically-derived carotenoid in birds. Comp Biochem Phys B . 132: 811– 818. Google Scholar CrossRef Search ADS   McKinnon JS, McPhail JD. 1996. Male aggression and colour in divergent populations of the threespine stickleback: experiments with animations. Can J Zool . 74: 1727– 1733. Google Scholar CrossRef Search ADS   Milinski M, Bakker TCM. 1990. Female sticklebacks use male coloration in mate choice and hence avoid parasitized males. Nature . 344: 330– 333. Google Scholar CrossRef Search ADS   Morrongiello JR, Bond NR, Crook DA, Wong BB. 2010. Nuptial coloration varies with ambient light environment in a freshwater fish. J Evol Biol . 23: 2718– 2725. Google Scholar CrossRef Search ADS PubMed  Olson VA, Owens IP. 1998. Costly sexual signals: are carotenoids rare, risky or required? Trends Ecol Evol . 13: 510– 514. Google Scholar CrossRef Search ADS PubMed  Östlund-Nilsson S. 2007. Reproductive behaviour in the three-spined stickleback. In: Östlund-Nilsson S, Mayer I, Huntingford FA, editors. Biology of the three-spined stickleback . Boca Raton (FL): CRC Press. p. 157– 178. Pike TW. 2017. Data from: discrimination of signal carotenoid content using multidimensional chromatic information. Dryad Digital Repository . http://dx.doi.org/10.5061/dryad.rd7jt Pike TW, Bjerkeng B, Blount JD, Lindstrom J, Metcalfe NB. 2011. How integument colour reflects its carotenoid content: a stickleback’s perspective. Funct Ecol . 25: 297– 304. Google Scholar CrossRef Search ADS   Pike TW, Blount JD, Bjerkeng B, Lindström J, Metcalfe NB. 2007. Carotenoids, oxidative stress and female mating preference for longer lived males. Proc Biol Sci . 274: 1591– 1596. Google Scholar CrossRef Search ADS PubMed  Pike TW, Blount JD, Lindstrom J, Metcalfe NB. 2007. Dietary carotenoid availability influences a male’s ability to provide parental care. Behav Ecol . 18: 1100– 1105. Google Scholar CrossRef Search ADS   Pryke SR, Andersson S. 2003. Carotenoid-based status signalling in red-shouldered widowbirds (Euplectes axillaris): epaulet size and redness affect captive and territorial competition. Behav Ecol Sociobiol . 53: 393– 401. Rick IP, Bakker TC. 2008. Color signaling in conspicuous red sticklebacks: do ultraviolet signals surpass others? BMC Evol Biol . 8: 189. Google Scholar CrossRef Search ADS PubMed  Rick IP, Modarressie R, Bakker TCM. 2004. Male three-spined sticklebacks reflect in ultraviolet light. Behaviour . 141: 1531– 1541. Google Scholar CrossRef Search ADS   Romero-Diaz C, Richner H, Granado-Lorencio F, Tschirren B, Fitze PS. 2013. Independent sources of condition dependency and multiple pathways determine a composite trait: lessons from carotenoid-based plumage colouration. J Evol Biol . 26: 635– 646. Google Scholar CrossRef Search ADS PubMed  Rowe MP, Baube CL, Loew ER, Phillips JB. 2004. Optimal mechanisms for finding and selecting mates: how threespine stickleback (Gasterosteus aculeatus) should encode male throat colors. J Comp Physiol A . 190: 241– 256. Google Scholar CrossRef Search ADS   Smith VC, Pokorny J. 1995. Chromatic-discrimination axes, CRT phosphor spectra, and individual variation in color-vision. J Opt Soc Am A. 12: 27– 35. Stradi R, Hudon J, Celentano G, Pini E. 1998. Carotenoids in bird plumage: the complement of yellow and red pigments in true woodpeckers (Picinae). Comp Biochem Phys B . 120: 223– 230. Google Scholar CrossRef Search ADS   Svensson PA, Wong BBM. 2011. Carotenoid-based signals in behavioural ecology: a review. Behaviour . 148: 131– 189. Google Scholar CrossRef Search ADS   Tinbergen N. 1951. The study of instinct . London: Oxford University Press. Veen T, Ingley SJ, Cui RF, Simpson J, Asl MR, Zhang J, Butkowski T, Li W, Hash C, Johnson JBet al.   2013. anyFish: an open-source software to generate animated fish models for behavioural studies. Evol Ecol Res . 15: 361– 375. Vinkler M, Albrecht T. 2010. Carotenoid maintenance handicap and the physiology of carotenoid-based signalisation of health. Naturwissenschaften . 97: 19– 28. Google Scholar CrossRef Search ADS PubMed  von Schantz T, Bensch S, Grahn M, Hasselquist D, Wittzell H. 1999. Good genes, oxidative stress and condition-dependent sexual signals. Proc Biol Sci . 266: 1– 12. Google Scholar CrossRef Search ADS PubMed  Wedekind C, Meyer P, Frischknecht M, Niggli UA, Pfander H. 1998. Different carotenoids and potential information content of red coloration of male three-spined stickleback. J Chem Ecol . 24: 787– 801. Google Scholar CrossRef Search ADS   Westland S, Ripamonti C. 2004. Computational colour science using MATLAB. Chichester (UK): John Wiley. Wootton RJ. 1984. A functional biology of sticklebacks . Berkeley (CA): University of California Press. Google Scholar CrossRef Search ADS   Zbinden M, Largiader CR, Bakker TCM. 2004. Body size of virtual rivals affects ejaculate size in sticklebacks. Behav Ecol . 15: 137– 140. Google Scholar CrossRef Search ADS   © The Author(s) 2017. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

Behavioral EcologyOxford University Press

Published: Jan 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

Monthly Plan

  • Read unlimited articles
  • Personalized recommendations
  • No expiration
  • Print 20 pages per month
  • 20% off on PDF purchases
  • Organize your research
  • Get updates on your journals and topic searches

$49/month

Start Free Trial

14-day Free Trial

Best Deal — 39% off

Annual Plan

  • All the features of the Professional Plan, but for 39% off!
  • Billed annually
  • No expiration
  • For the normal price of 10 articles elsewhere, you get one full year of unlimited access to articles.

$588

$360/year

billed annually
Start Free Trial

14-day Free Trial