Temperature-dependent colour change is a function of sex and directionality of temperature shift in the eastern fence lizard (Sceloporus undulatus)

Transcription

1 Biological Journal of the Linnean Society, 2017, 120, With 5 figures. Temperature-dependent colour change is a function of sex and directionality of temperature shift in the eastern fence lizard (Sceloporus undulatus) BARRY P. STEPHENSON 1 *, NIKOLETT IH ASZ 2, DAVID C. BYRD 1, JOHN SWIERK 3 and LINDSEY SWIERK 4 1 Department of Biology, Mercer University, Macon, GA, 31027, USA 2 Department of Psychology, Mercer University, Macon, GA, 31207, USA 3 Department of Chemistry, The Pennsylvania State University, University Park, PA, 16802, USA 4 Department of Biology, Intercollege Graduate Program in Ecology and Center for Brain, Behavior and Cognition, The Pennsylvania State University, University Park, PA, 16802, USA Received 24 November 2015; revised 16 June 2016; accepted for publication 18 July 2016 Sexually dimorphic colour traits are widespread across taxa, but relatively little is known about how and why these features change with body temperature. To examine whether sex and directionality (warming vs. cooling) influence temperature-dependent colour change, we used spectrophotometry to characterize ventral and dorsal coloration in eastern fence lizards (Sceloporus undulatus), a species with temperature-dependent conspecific signals (ventral patches). In general, we found that skin colour (hue) in both sexes changed with temperature. Only ventral patch colour in males changed visibly to human observers; ventral patch hue decreased (green to blue shift) with an increase in body temperature. Male dorsal colour also changed with temperature, with cooler males exhibiting a longer-wavelength hue (red shift) than that exhibited by warmer males. Female dorsal hue changed in a manner paralleling that of males, and female ventral hue trended in the same direction: overall, warmer females tended to exhibit shorter-wavelength hues. However, the magnitude of male ventral patch colour change was highly dependent upon temperature sequence; the colourful badges of males that were progressively warmed changed from green to blue, whereas those of males that were cooled over the same time period did not change hue. Notably, only the coloration of male ventral patches was dependent on the directionality of temperature change; no other measured skin surfaces in either sex exhibited similar temperature-directionality dependence. These results are considered in the context of models for badge colour production in Sceloporus, as well as in hypotheses for the functional significance of ventral patch coloration The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 2017, 00, 120, KEYWORDS: body temperature sexual dimorphism spectrophotometry spiny lizard. INTRODUCTION To cope with changes in their environments, many species possess the ability to alter body coloration reversibly in response to abiotic or social factors. Rapid body colour change has been documented in almost all animal taxa (e.g. O Connor, Metcalfe & Taylor, 1999; Langridge, Broom & Osorio, 2007; Ries et al., 2008; Vroonen et al., 2012; Umbers et al., 2014) and is often used by researchers as a window *Corresponding author: stephenson_bp@mercer.edu Current address: School of Forestry and Environmental Studies, Yale University, New Haven, CT, 06511, USA into an individual s physiological state (e.g. Yang et al., 2001; Greenberg, 2002; Plavicki, Yang & Wilczynski, 2004) because of the relationship of body colour to stress level, reproductive receptivity, or social dominance. Nevertheless, the factors eliciting rapid colour change in most species are not well understood (e.g. Umbers et al., 2014). In particular, while understanding of the physiology and function of sexual dichromatism is well developed (reviewed in Shine & Madsen, 1994; Badyaev & Hill, 2003; Bell & Zamudio, 2012), how and why the sexes differ in rapid colour change is largely unexplored. Furthermore, how gradual changes in environmental factors (such as those actually experienced by an animal in 396

2 the field) influence rapid colour change has largely been founded on experiments providing all-or-nothing exposure to stimuli [e.g. temperature (Silbiger & Munguia, 2008) and background colour (Stevens, Lown & Denton, 2014)]. Lizards are exceptionally well suited to studies designed to elucidate the drivers and functions of rapid colour change, as many species produce dramatic colour changes in response to external factors. Such changes can be directly induced by social interactions, such as the striking colours exhibited by many species of chameleons in response to courtship or aggression (Stuart-Fox & Moussalli, 2008) or the dorsal darkening observed in anoles responding to changes in dominance hierarchies (Jenssen, Greenberg & Hovde, 1995). Many lizards exhibit temperature-dependent colour change; such species tend to be darker in the morning to facilitate rapid warming and lighter in the heat of the day to prevent overheating (Norris, 1967; Walton & Bennett, 1993; Rosenblum, 2005). Given the relationship between internal body temperature and physiology fundamental to ectotherms, it has been suggested that temperature-dependent colour change could also represent another signal of performance capacity (Langkilde & Boronow, 2012). The eastern fence lizard, Sceloporus undulatus (Bosc and Daudin), is an excellent model of conspicuous coloration and rapid colour change. This species is a common and widespread reptile in the eastern and central USA (Conant & Collins, 1998). In Georgia, this species is especially abundant in upland, dry forested habitats (Jensen et al., 2008). Like most others within its genus, S. undulatus is characterized by the presence of sexually dimorphic blue colour patches on the throat and abdomen, features revealed to conspecifics during courtship and aggressive interactions (Wiens, 1999). Among Sceloporus lizards, variation in the colour or size of ventral patches has been proposed to advertise sex (Cooper & Burns, 1987), reproductive state (Weiss, 2006), fighting ability (Rand, 1988), parasite load (Ressel & Schall, 1989; Calisi & Hews, 2007), mate quality (Bastiaans et al., 2014), and species identity (Quinn & Hews, 2000). However, Langkilde & Boronow (2010) and Quinn (2001) found little support for the hypothesis that variation in male throat and abdominal colour patches (at constant temperature) correlates with morphological traits linked to fitness specifically in S. undulatus. This contrast leaves open the question of the social function of ventral patches in this species. A notable feature of this conspicuous signal is its striking temperature-dependent colour change (Langkilde & Boronow, 2012). At warmer temperatures (29 C), these patches exhibit an intense COLOUR CHANGE IN FENCE LIZARDS 397 metallic blue colour, whereas at temperatures only a few degrees cooler (25 C), male ventral colour is green (Langkilde & Boronow, 2012). To our knowledge, strong change in sexually dimorphic ventral colour traits as a function of temperature has not been formally described in other species of Sceloporus (but see Ressel, 1988). Consequently, the temperature dependence of ventral patches may offer a clue to their social function in this species. As a first step to addressing these issues, we assessed colour variation in several sexually dimorphic colour traits in adult S. undulatus using spectrophotometry at four ecologically relevant temperatures. Although other studies have used spectrophotometry to describe body colour expression in S. undulatus, temperature at measurement was not presented (Stoehr & McGraw, 2001; Rosenblum, 2005) or within-species sample sizes were too small to be informative (Hutchison & Larimer, 1960). Crucially, none of these studies involved directly manipulating temperature as a variable on colour expression, as in Langkilde & Boronow (2012), in which data were collected via photography and could not be assessed for critical features such as ultraviolet (UV) colour (Stoehr & McGraw, 2001). Based on the idea that rapid colour change may be an honest signal in this species because of its association with endurance and physiological capacity, we predicted that spectral shifts would correspond to body temperature shifts in a manner similar to that described in Langkilde & Boronow (2012): warmer (i.e. more physiologically active) lizards would display more conspicuous blue/uv ventral patches (potentially serving as a signal of fighting ability to conspecifics) and lighter dorsal coloration (to modulate thermoregulation more effectively). These patterns may be sex-dependent, as in other taxa (e.g. Uca crabs; Silbiger & Munguia, 2008); male S. undulatus expend greater energy and remain exposed to maintain fitness-enhancing territories (Haenel, Smith & John-Alder, 2003), and females are generally not observed to display their ventral surface during competitive interactions. Furthermore, we predicted that the directionality of temperature change (heating vs. cooling) relates to the colour adopted by S. undulatus. As internal body temperature can be maintained under cooling conditions (Bartholomew & Tucker, 1963; Bartholomew & Lasiewski, 1965), we expected that rapid dorsal colour change occurs in warming (dark-to-light dorsa), but not cooling (light-to-dark dorsa), conditions. For similar reasons, males (but not females) may maintain conspicuous ventral coloration in cooling conditions. We place our results in context of the potential physiological mechanisms and social functions of body colour in ectotherms with rapid colour change.

3 398 B. P. STEPHENSON ET AL. MATERIAL AND METHODS COLLECTION AND CAPTIVE MAINTENANCE Lizards were collected between 28 April and 25 June, 2013, at Dauset Trails Nature Center, Jackson, Georgia, USA (33.14 N, W). The Georgia Department of Natural Resources provided a collecting permit (#29-WJH-13-74), and all research was approved by the Mercer University Animal Care and Use Committee (#A ). We captured lizards using a telescoping fishing pole with a slipknot at the end, and placed them into individual cloth bags for transfer, within a few hours, to animal care facilities at Mercer University. Upon arrival at Mercer, we placed each lizard into an individual glass terrarium (61 cm 9 32 cm 9 42 cm) containing a full-spectrum light source, vertical wood structure for perching, wood chip bedding, rocks and bark for shelter, and a small bowl of water. Lizards were kept visually isolated from each other by cardboard partitions placed between adjacent terraria. A 12-light:12- h dark light cycle was fixed for the duration of the experiment, and room temperature was maintained at between 21.3 C and 29.8 C. All lizards were allowed at least 1 day to recover from the stress of field capture and transport before use in experiments (see Temperature treatments). During captivity, lizards were given access to water continuously, but provided with food (two to four crickets every 2 days) only following collection of colour and morphological data, to preclude any possible interference of temporary diet with spectral properties of colour patches (Olson & Owens, 1998). Following no longer than 6 days in captivity, each lizard was released at its respective capture site. Before release, we applied a small white dot of nontoxic paint (Grumbacher, Chartpak, Inc., Leeds, MA, USA) to the base of the tail to help prevent recapture of previously measured lizards (Simon & Bissinger, 1983). As exposure to water and natural skin shedding (ecdysis) remove paint over time, we further minimized the potential for multiple captures by avoiding areas extensively searched earlier in the investigation. Terraria were thoroughly cleaned between use with hot soap and water, and chip bedding and drinking water were replaced. All captured male lizards exhibiting well-developed abdominal and throat patches of the venter were considered as sexually mature and were retained for use in experiments [snout-vent length (SVL): X standard error of the mean = mm; range = mm; N = 28]; a few, very small, males that did not meet these criteria were released. All adult females used in experiments (SVL: X standard error of the mean = ; range = mm; N = 12) were palpated for signs of oviductal eggs; any female interpreted to be gravid was immediately released to prevent oviposition in the laboratory. We note that females in some Sceloporus species exhibit sexually dimorphic colour traits that covary in expression with reproductive state [e.g. Sceloporus virgatus (Weiss, 2006) and Sceloporus pyrocephalus (Calisi & Hews, 2007)]. As females interpreted to be gravid when captured were selectively removed from this analysis, we cannot rule out this possibility in S. undulatus. TEMPERATURE TREATMENTS We measured lizards for body colour at each of four temperature treatments: 24 C, 28 C, 32 C, and 36 C. This extended the temperature range used by Langkilde & Boronow (2012) to include estimates of body temperature (T b ) in the field (~35 C; Andrews, 1998) during periods of peak activity, as well as selected temperatures (T sel ) in the laboratory (Ballinger, Hawker & Sexton, 1969; Crowley, 1987), but below critical temperatures where performance may be compromised (Ballinger et al., 1969). Temperature treatments proceeded linearly, either from lowest to highest or from highest to lowest, with the ordering of treatment sequence alternated between lizards of a given sex with respect to their capture sequence in the field. For a given sex, the sequence of the first lizard captured was determined by a coin-flip. Before the start of a given sequence of temperature treatments, we removed the selected lizard from its terrarium and placed it within a clean, dry transparent plastic box (15 cm 9 8 cm 9 13 cm) containing myriad air holes for ventilation. The box was then placed inside a portable incubator (ExoTerra PT2199) and each lizard was allowed at least 15 min to acclimatize to the initial temperature setting (i.e. either 24 C or 36 C). A thermometer fixed to each plastic box containing a lizard provided confirmation of the temperature experienced within the box, as the external temperature reading reported by the incubator was subject to greater variability (B.P. Stephenson, pers. observ.). After interior temperature readings indicated that air temperature in the incubator was within 1 C of the desired temperature, we removed the box from the incubator and opened the lid. To minimize handling before capturing colour data, we used an infrared thermometer (Ryobi Tek4 RP4030: 0.1 C) to measure the temperature of the right side of the dorsum of the subject, about 20 cm from the thermometer (Hare, Whitworth & Cree, 2007; Carretero, 2012). Up to three temperature readings were collected per lizard; although all three temperature readings were, on average, within 2 C of the intended temperature, we used 1 C as our minimum accepted

4 deviation from the intended temperature. If no reading fell within 1 C of the intended temperature, we returned the lizard to the incubator. After allowing it to warm or cool for a few more minutes, we removed the lizard from the incubator again and remeasured it for temperature, as described above. There was no difference between males and females in the time spent in the incubator to reach the initial selected temperature and between successive measurements [the following values are given as mean standard error of the mean (range) average temperature adjustment period: males, 42 2 (range: 22 61) min; females, 42 2 (range: 28 55) min; t 38 = 0.22; P = 0.82; see Tables S1 and S2]. There was also no difference in the total time spent in the incubator between males as a function of treatment sequence [warming: (range: ) min; cooling: (range: ) min; Kruskal Wallis Z = 1.2; P = 0.22] or in the total time that females spent in the incubator as a function of treatment sequence [warming: (range: ) min; cooling: (range: ) min; Kruskal Wallis Z = 0.6; P = 0.52]. If body temperature was within the acceptable range of the given temperature treatment, the lizard was removed from its plastic box, photographed (both dorsal and ventral perspectives) against a white background, and measured for body colour (see Spectrophotometry). Following collection of the first set of colour data, we returned each lizard to the incubator and raised or lowered the temperature to the next temperature treatment accordingly. We repeated this process as described above for all four temperature treatments. Following collection of the final set of colour data, we measured lizards for various morphometric variables as part of a separate study, and then returned them to their individual glass terraria, providing food and water. Lizards were given at least 1 day to recover from any stress associated with experimental procedures before being released at their sites of capture. SPECTROPHOTOMETRY We used spectrophotometry to measure body colour in lizards following attainment of targeted body temperature. A pulsed xenon light source (PX-2: Ocean Optics, Inc.) was used to illuminate lizard skin; light was transmitted by a probe contained within a black probe holder fixed at a 45 angle. Reflectance was then transmitted, via a separate channel in the same probe, to an Ocean Optics USB4000 spectrometer for data processing and the results were analysed relative to a 99% white standard using SpectraSuite software (Ocean Optics: Dunedin, FL, USA). Spectral data were collected in the presence of illuminated COLOUR CHANGE IN FENCE LIZARDS 399 ceiling lights; however, we anticipate little or no interference with this illumination and spectral data collection as a result of the shielding effect of the probe holder and the diffuse nature of fluorescent light. During each of four spectral measurement sessions, we measured each male lizard three times on the right blue green ventral abdominal patch (Fig. 1A) and on the right side of the dorsum. Females were measured in the same areas on the right side of the dorsum and venter (Fig. 1B, C), but generally lacked a blue or green ventral abdominal patch. We calculated the wavelength of peak reflectance (hue: k Rmax ) within the 300- to 700-nm region (i.e. the range of lizard visual sensitivity; Loew et al., 2002) for each individual spectrum. We then characterized spectral reflectance for dorsal and ventral abdominal coloration for each lizard at a given temperature by determining the single spectrum associated with the median hue value. Individual spectra were removed before analysis (and determination of median hue for a given temperature) under any of three conditions: (1) two consecutive spectra were found to be identical, indicating failure to capture a new spectrum, within a series, correctly; (2) the hue for a given spectrum was associated with peak reflectance of < 10%; and (3) the shape of the spectrum was strongly suggestive of inappropriate probe placement. Although this latter criterion is subjective, only those spectra that were highly divergent (i.e. strongly dissimilar in shape and brightness from other spectra of the same measured feature; see Fig. S1 for a representative example) were excluded. When only one spectrum (out of three) was removed Figure 1. Representative examples of colour badges of adult Sceloporus undulatus measured in this study. Red ovals denote areas measured. A, right male ventral patch. B, right female venter. C, right female dorsum.

5 400 B. P. STEPHENSON ET AL. from a series, the spectrum used in subsequent analyses was determined by random selection. From each selected spectrum, we also calculated UV chroma (R /R ) and brightness (R ), as described elsewhere (Montgomerie, 2006; Thorogood et al., 2008). Spectra exclusion criteria were fixed before analysis and closely conform to those given previously in Stephenson (2010). STATISTICAL ANALYSES We used repeated-measures analysis of variance (ANOVA) to test for differences in spectral reflectance of each of three raw colour variables (hue, UV chroma, and brightness) under each temperature treatment, with sequence of temperature treatment (increasing or decreasing) serving as the betweensubjects grouping variable. When Mauchly s test of sphericity was significant, we used the Greenhouse Geisser estimate of e when e < 0.75, and the Huynh Feldt estimate of e when e > 0.75 following the recommendations of Field (2009), with corresponding adjustments to the degrees of freedom for such tests. We natural-log transformed all colour variables to provide better conformity to assumptions of normality. We provide effect sizes for all tests of repeatedmeasures ANOVA (calculated according to Lakens, 2013) in Table S3. We used IBM SPSS v. 21 and JMP v for statistical analyses. All tests were two-tailed, and significance was determined at a = Means reported as X standard error of the mean. RESULTS Between the two treatment sequence groups, there was no difference in SVL of either males (warming: X = mm; cooling: X = mm; Kruskal Wallis Z = 0.05; P = 0.96) or females (warming: X = mm; cooling: X = mm; Kruskal Wallis Z = 0.40; P = 0.69). MALE VENTRAL PATCH COLOUR In accordance with screening procedures described previously, we removed 27 (8.0%) of 336 spectra because of concerns of flawed measurement. Across all temperature treatments, male ventral patches of the abdomen showed a single major peak centred around nm (Fig. 2A, B; see also Fig. S2). Brightness in the UV region ( nm) was low ( X = 6.2%) across all temperatures. Our results indicate a significant within-subjects effect of temperature on ventral patch hue (F 3,78 = 9.72, P < 0.001). Males that were warmer Figure 2. Colour change in adult male Sceloporus undulatus. A, male ventral and dorsal abdominal coloration from an individual male in the warming temperature treatment sequence. B, mean ventral patch reflectance spectra for each temperature treatment (N = 28). C, mean dorsal reflectance spectra for each temperature treatment (N = 28). Averaged spectra (Avg) are plotted as medians of 10-nm bins for clarity. had ventral patches that were bluer (i.e. had shorter hue) than males that were cooler. There was also a significant interaction of temperature and sequence on hue (F 3,78 = 5.63, P = 0.002; Fig. 3A): males in the warming group exhibited a smooth decline in ventral patch hue from 540 to 470 nm, consistent with a transition from green to blue. Males in the

6 COLOUR CHANGE IN FENCE LIZARDS 401 Figure 3. Change in male ventral (A C) and dorsal (D F) abdominal coloration with temperature. A, D, hue; B, E, UV chroma; C, F, brightness. Grey bars indicate warming treatment (N = 14) and white bars indicate cooling treatment (N = 14). Data are presented as mean standard error of the mean. cooling group did not exhibit a green ventral patch by the end of the temperature sequence, but instead remained blue. For these males, peak mean hue was observed at 32 C. There was no significant betweensubjects effect of sequence on hue (F 1,26 = 0.04, P = 0.84). Temperature, sequence, and their interaction did not affect either UV chroma or brightness (Fig. 3B, C; all P > 0.08). MALE DORSAL COLOUR We removed two (0.6%) of 336 spectra before analysis. Across all treatments, male dorsal patches showed a single major peak centred around nm (Fig. 2C). Brightness in the UV region was low ( X = 5.4%) across all temperatures. Temperature affected dorsal hue such that warmer males exhibited dorsal colour with a shorter peak wavelength compared with that exhibited by cooler males (F 3,78 = 38.06, P < 0.001). There was no effect of sequence (F 1,26 = 0.06, P = 0.81), or the interaction of temperature and sequence (F 3,78 = 1.11, P = 0.35; Fig. 3D), on hue. There was an interactive effect of temperature and sequence (F 3,78 = 3.09, P = 0.03; Fig. 3E) on UV chroma, which decreased with higher temperature in the warming group but increased in the cooling group. However, temperature and sequence alone did not affect UV chroma (temperature: F 3,78 = 0.38, P = 0.76; sequence: F 1,26 = 0.20, P = 0.65). There was no significant effect of temperature, sequence, or their interaction on brightness (Fig. 3F; all P > 0.65).

7 402 B. P. STEPHENSON ET AL. FEMALE VENTRAL COLOUR We rejected two (1.4%) of 144 spectra before analysis on suspicion of flawed measurement. Female ventral colour of the abdomen differed substantially from that of males, with relatively high reflectance across all wavelengths, and a single peak between 640 and 670 nm (Fig. 4A, B; see also Table S3). Mean UV reflectance ( X= 21.6%) was higher in females than in males across all treatments. There was no effect of temperature, sequence, or their interaction on female abdominal hue, UV chroma, or brightness (Fig. 5A C; all P>0.13). FEMALE DORSAL COLOUR We removed one (0.7%) of 144 spectra before analysis. Female dorsal colour resembled that of males, with low reflectance across all wavelengths and a single peak between 690 and 700 nm (Fig. 4C). Mean UV reflectance across all treatments ( X= 5.9%) was similar to that observed for male dorsal colour. Temperature affected female dorsal hue (F 1.2,12.5 = 5.60, P = 0.029), such that warmer females exhibited dorsal colour with a shorter hue than did cooler females, but neither sequence nor the interaction of temperature and sequence affected hue (sequence: F 1,10 = 0.24, P = 0.63; interaction: F 1.2,12.5 = 0.38, P = 0.60, Fig. 5D). Likewise, temperature affected female dorsal UV chroma (F 3,30 = 6.35, P = 0.002); cooler females had higher dorsal UV chroma than did warmer females. There was no effect of sequence or the interaction of temperature and sequence on dorsal UV chroma (sequence: F 1,10 = 0.38, P = 0.55; interaction: F 3,30 = 0.93, P = 0.44; Fig. 5E). There was a significant effect of temperature on brightness (F 3,30 = 2.93, P = 0.049); females at 32 C were brighter than females at 28 C only. However, there was no effect of sequence or the interaction of temperature and sequence on female dorsal brightness (sequence: F 1,10 = 1.12, P = 0.31; interaction: F 3,30 = 1.03, P = 0.39; Fig. 5F). DISCUSSION Studies of animal colour have generated substantial insights into our understanding of sexual selection (Cuthill et al., 1999; Hunt et al., 1999), honest signalling (Hill, 1990; Grether, 2000), predator prey interactions [e.g. aposematism (Brodie & Janzen, 1995) and camouflage (Stuart-Fox et al., 2003)], and thermoregulation (Geen & Johnston, 2014). Many lizards exhibit dramatic colour change with temperature (Norris, 1967; Walton & Bennett, 1993), a Figure 4. Colour change in adult female Sceloporus undulatus. A, female ventral and dorsal abdominal coloration from an individual female in the warming temperature treatment sequence. B, mean ventral reflectance spectra for each temperature treatment (N = 12). C, mean dorsal reflectance spectra for each temperature treatment (N = 12). Averaged spectra (Avg) plotted as medians of 10-nm bins for clarity. phenomenon usually interpreted as an adaptation to thermoregulation. Less is known about the extent and possible functional significance of temperature dependence of sexually dimorphic colour badges (Langkilde & Boronow, 2012), especially those of the ventral surface which are ordinarily not directly

8 COLOUR CHANGE IN FENCE LIZARDS 403 Figure 5. Change in female ventral (A C) and dorsal (D F) abdominal colour with temperature. A and D, hue; B and E, UV chroma; C and F, brightness. Grey bars indicate warming treatment (N = 6) and white bars indicate cooling treatment (N = 6). Data are presented as mean standard error of the mean. exposed to solar radiation (Smith et al., 2016). Here, we found that rapid colour change in both dorsal and ventral colour expression in S. undulatus is linked to body temperature. Moreover, such changes were observed primarily in hue rather than in brightness (or UV chroma), suggesting that any functional basis for colour change is unrelated to thermoregulation. MALE BODY COLOUR Ventral abdominal patches of males were blue when warm and green when cool (Fig. 2A), corresponding to earlier findings of Langkilde & Boronow (2012) from other S. undulatus populations. We found that this colour change was caused by a decrease in the wavelength of peak reflectance with increasing temperature (i.e. hue: k Rmax 36 C = 492 nm; k Rmax 24 C = 525 nm), whereas ventral patch UV chroma and brightness did not change with temperature (Fig. 3A C, Table 1). A notable finding of our study was the effect of temperature sequence (warming vs. cooling) on ventral patch colour change; males that were progressively warmed from a cool initial temperature exhibited the conspicuous change from green to blue, but males that were progressively cooled after initial warming did not (Fig. 3A), indicating an asymmetry in the rate of ventral patch colour change. Interestingly, this effect was not observed in male dorsal coloration or in females, which suggests that there may be fundamental differences in the physiology of male colour badges, a functional cause for colour badge temperature

9 404 B. P. STEPHENSON ET AL. Table 1. Summary of male ventral and dorsal abdominal colour at each of four temperature treatments 24 C 28 C 32 C 36 C Variable X Range X Range X Range X Range Venter Hue (nm) UV chroma (%) Brightness (%) Dorsum Hue (nm) UV chroma (%) Brightness (%) Values are given as mean standard error of the mean and range (as indicated). N = 28 for all treatments. asymmetry, or both (see Possible functional significance of colour change). It is notable that blue badges in cooling group males do eventually return to green (i.e. by the following day; B.P. Stephenson and L. Swierk, pers. observ.), suggesting that the blue-to-green transition occurs on a much slower timescale than the reverse and may be dependent on additional physiological or environmental factors. Shifts in the hue of the male dorsum (Fig. 3D, E; Table 1) largely mirrored the overall magnitude and direction of ventral patch colour change. Although this shift was not nearly as conspicuous as ventral patch colour change to human observers, cooler males exhibited a significantly longer wavelength dorsal hue (k Rmax 24 C = 688 nm) compared with that of warmer males (k Rmax 36 C = 668 nm). We also found that warming males showed increased UV chroma with increasing temperature, whereas cooling males showed increased UV chroma with decreasing temperature. Because increased UV chroma corresponds to increased length of time spent in the experimental trial (in both the cooling and warming treatments), one possibility is that increases in dorsal UV chroma might be a by-product of a generalized response to stress (i.e. as a function of prolonged handling and confinement during temperature measurements). However, we did not observe this pattern in any other measured feature in either sex, and thus any possible biological relevance to this pattern is unclear. FEMALE BODY COLOUR Female ventral abdominal colour differed markedly from that of males, characterized by an overall pale coloration which was brighter than that of males and a lack of visible patches (Figs 1B, 4B; Tables 1, 2, S4), a sexual dimorphism seen in some other lizards (e.g. LeBas & Marshall, 2000). There was a trend (although non-significant) for ventral colour hue in females to decrease with increasing temperature. As we found evidence of temperature-dependence for male and female dorsal colour and male ventral patch colour, the integumentary basis for a role of temperature on skin colour is likely to be a general feature of the skin of both sexes in this species. Table 2. Summary of female ventral and dorsal abdominal colour at each of four temperature treatments Variable 24 C 28 C 32 C 36 C X Range X Range X Range X Range Venter Hue (nm) UV chroma (%) Brightness (%) Dorsum Hue (nm) UV chroma (%) Brightness (%) Values are given as mean standard error of the mean and range (as indicated). N = 12 for all treatments.

10 COLOUR CHANGE IN FENCE LIZARDS 405 In other populations of S. undulatus, many females exhibit conspicuous green or blue ventral abdominal and throat patches, although these tend to be much less prominent than those of males (Swierk & Langkilde, 2013a). Some females in our population also exhibit throat (see Fig 1B) and ventral (B.P. Stephenson, pers. observ.) patches, but their colour expression is paler and greener than that reported by Swierk & Langkilde (2013a), and no female exhibited intense blue patches of the kind seen in warm adult males. It is unknown whether females with male-like ventral patches (e.g. Swierk & Langkilde, 2013a; Gilbert & Lattanzio, 2015) would exhibit temperature-dependent colour-change patterns similar to those of males, but investigation into this question would shed light on the proximate causes and function of colour change in S. undulatus. BRIGHTNESS We found little evidence for an effect of temperature on integumentary brightness in S. undulatus, suggesting that skin brightness is insensitive to ambient temperature, at least over the range of temperatures used in this study. This is surprising, given that many squamate reptiles regulate their body temperature by modulating skin reflectance (Norris, 1967). For example, short-term changes in dorsal brightness with respect to temperature were substantial for lizards from the blanched S. undulatus populations at White Sands, NM, although, interestingly, not for those from the adjacent dark soil populations (Rosenblum, 2005) which are much more similar in appearance to the population we used in our study. Alternatively, integumentary brightness might be affected by wavelengths not measured in this study. In support of this point, Walton & Bennett (1993) found that brightness in the near-infrared region was strongly affected by temperature in two of three chameleon species, although such changes also extended into the visible spectrum. A third possibility is that any such effect of temperature on brightness requires longer periods of time to become apparent than that employed in our experimental regime. Given the importance of changes in brightness to desert reptiles (Norris, 1967), including some close relatives of S. undulatus, we cannot rule out this possibility entirely. However, the lizards in our study experienced biologically relevant increases and decreases in temperature across a 2-h period of time, and it is unclear how a change in brightness, in response to our morning-warming or midday-cooling simulations, would not have been detected in our study. PROXIMATE BASIS OF TEMPERATURE-DEPENDENT COLOUR CHANGE Reptilian colour change is thought to be caused primarily by changes in skin brightness (i.e. rather than hue or chroma) arising from shifts in the dispersion of melanin within dermal melanophores (Cooper & Greenberg, 1992); the more dispersed the melanin within the melanophore layer, the darker (i.e. less reflective) the integument. However, most lizards also possess layers of cells superficial to the melanophore layer (e.g. Morrison, Rand & Frost-Mason, 1995), with the consequence that any light reflected from deeper tissues (i.e. not absorbed by melanin) represents only one component of the total reflected from the body surface (Cooper & Greenberg, 1992; see also Hews & Quinn, 2003). Within these layers are iridophores, which contain platelets of crystalline guanine (stored within iridosomes) that selectively reflect short wavelength light by Rayleigh scattering (Hews & Quinn, 2003; Olsson, Stuart-Fox & Ballena, 2013). Morrison, Sherbrooke & Frost-Mason (1996) suggest that decreased spacing between adjacent iridosomes reduces the scattering of long-wavelength radiation, contributing to changes in hue such as the increased blueness of blue abdominal tissue in Urosaurus ornatus. These observed shifts were temperature dependent, with warmer temperatures resulting in iridosomes packed more closely together (Morrison et al., 1996; see also Teyssier et al., 2015). Given the close evolutionary relationship of Urosaurus and Sceloporus (Wiens et al., 2010), the iridosome shifts are likely to contribute to the abdominal hue shift first observed by Langkilde & Boronow (2012) and quantified in our study. The observation that male colour did not revert to green from blue with progressive cooling over the same period of time as in males subjected to progressive warming warrants further consideration. From a proximate standpoint, this suggests that the transition to green from blue either takes longer to accomplish than transitions to blue from green, or that some other mechanism responsible for the colour production has not been identified. However, it should be noted that endogenous heat production and circulatory heat transport allow lizards to maintain optimal internal temperatures for longer periods under cooling conditions (Bartholomew & Tucker, 1963; Bartholomew & Lasiewski, 1965), resulting in lizards heating more rapidly than they cool (Kour & Hutchison, 1970; see also Rice & Bradshaw, 1980). We assessed body temperature using a well-established technique (infrared laser thermometer; Berg, Theisinger & Dausmann, 2015), although skin temperature alone may be an incomplete picture of bodycolour regulation (see Taylor, Tipton & Kenny,

11 406 B. P. STEPHENSON ET AL. 2014). Consequently, future studies comparing in vivo and in vitro temperature-dependent colour change in abdominal tissue would help to elucidate the proximate causes of this heating/cooling colourchange asymmetry. POSSIBLE FUNCTIONAL SIGNIFICANCE OF COLOR CHANGE Overall, we found that both dorsal and ventral coloration in males and females changed in response to temperature; moreover, the spectral variable most strongly affected by temperature was hue. Higher temperatures generally resulted in lower (i.e. bluer) hues for both dorsal and ventral coloration, consistent with the model of colour change proposed by Morrison et al. (1996). The most dramatic change in coloration, both to the human eye and in terms of spectral differences, occurred on the male ventral patches. The functional significance of male ventral colour change, if any, has yet to be identified. One possibility is that this temperature-dependent colour change is a by-product of integumentary colour production in this species and serves no function in and of itself. Alternatively, temperature dependence of body colour could serve as an honest signal of physiological condition. Colder lizards are physiologically compromised compared with warmer lizards near their thermal optimum (around 33 C for populations in the south-eastern USA; Angilletta, 2001); consequently, the bluer colours seen in warmer lizards could honestly advertise male physiological capacity (Langkilde & Boronow, 2012; Beal, Lattanzio & Miles, 2014). As Sceloporus lizards approach warmer and more optimal temperatures, they have greater endurance and sprint velocity (Marsh & Bennett, 1986; Angilletta, Hill & Robson, 2002; Beal et al., 2014) and increased metabolic rate (Angilletta, 2001), which has been shown to permit greater capacity to defend territories (Marler et al., 1995) and maintain dominance (Metcalfe, Taylor & Thorpe, 1995) in ectotherms. On the other hand, lizards maintaining cooler body temperatures have the advantage of reduced energy expenditure (Grant, 1990; Angilletta et al., 2002) and tend to exhibit more aggressive behaviours, probably because of their reduced abilities to flee (Mautz, Daniels & Bennett, 1992). As such, if male abdominal colour change originated as a by-product of integumentary colour production, it may have been later co-opted to serve as a dynamic signal in male male contests. Male S. undulatus are known to alter their competitive behaviours depending on both the display type and relative sizes of other males (Swierk & Langkilde, 2013b). Like behaviour, body-temperature changes on relatively short timescales has substantial implications for competition. Males may therefore benefit by using an obvious signal of temperature (i.e. ventral patch colour) to inform their tactics during male male contests. We note that other animal signalling systems can be directly affected by ambient temperature, especially acoustic signals (e.g. tree frogs: Gerhardt, 1978; crickets: Doherty, 1985; see also Gillooly & Ophir, 2010). As performance capacity in ectotherms is often closely linked to temperature (Angilletta et al., 2002), if colour signal expression is also correlated with fitnessrelated performance measures, and if receivers can detect such changes in colour expression (e.g. Endler, 1990), then temperature-dependent colour change could form the basis of a stable signalling system in this species and others (Langkilde & Boronow, 2012). Visual modelling studies that evaluate the extent to which colour change is perceptible by intended receivers (e.g. Smith et al., 2016), and experiments that evaluate receiver responses to manipulated colour badges (e.g. Thompson & Moore, 1991) will be beneficial in this regard. In addition, some females in our population bore light-green ventral patches that resemble those of very pale (and thermally cool) males (B.P. Stephenson, pers. observ.). No females were observed with blue badges, although some females bear blue badges in other populations (Swierk & Langkilde, 2013a). If adult males are more likely to, for example, escalate contests with adult lizards with bright blue patches, then male S. undulatus that resemble females may be less likely to receive aggression from neighbouring males, thus sparing cold, green males contests with warm, blue rivals that they would likely lose. Male S. undulatus are known to use the presence or absence of blue badges as the primary indicator of an individual s sex (Cooper & Burns, 1987); an approaching individual painted with blue patches is treated aggressively, regardless of its actual sex, whereas white-painted individuals are initially courted. However, whether green badges are similarly used as sex identifiers is unknown. Future studies should seek to illuminate this complex interaction of body colour, temperature, and fitness to broaden our understanding of the social functions of rapid colour change across taxa. ACKNOWLEDGEMENTS We thank Dauset Trails Nature Center for allowing us to collect lizards, and providing additional logistic support during this study. William J. Jenkins provided advice on statistics, and three anonymous reviewers provided helpful comments that strengthened the manuscript. The Biology Department at

12 Mercer University provided financial support to BPS for this research. The National Science Foundation provided additional support to LS (grant no. DGE ). Portions of the research were presented at the 76th meeting of the Association of Southeastern Biologists, in Chattanooga, TN in April, REFERENCES Andrews RM Geographic variation in field body temperature in Sceloporus lizards. Journal of Thermal Biology 23: Angilletta MJ Jr Thermal and physiological constraints on energy assimilation in a widespread lizard (Sceloporus undulatus). Ecology 82: Angilletta MJ Jr, Hill T, Robson MA Is physiological performance optimized by thermoregulatory behavior?: a case study of the eastern fence lizard, Sceloporus undulatus. Journal of Thermal Biology 27: Badyaev AV, Hill GE Avian sexual dichromatism in relation to phylogeny and ecology. Annual Review of Ecology, Evolution, and Systematics 34: Ballinger RE, Hawker J, Sexton OJ The effect of photoperiod acclimation on the thermoregulation of the lizard Sceloporus undulatus. Journal of Experimental Zoology 171: Bartholomew GA, Lasiewski RC Heating and cooling rates, heart rate, and simulated diving in the Galapagos marine iguana. Comparative Biochemistry and Physiology 16: Bartholomew GA, Tucker VA Control of changes in body temperature, metabolism, and circulation by the agamid lizard, Amphibolurus barbatus. Physiological Zoology 36: Bastiaans E, Bastiaans MJ, Morinaga G, Casta~neda Gaytan JG, Marshall JC, Bane B, Mendez de la Cruz F, Sinervo B Female preference for sympatric vs. allopatric male throat color morphs in the mesquite lizard (Sceloporus grammicus) species complex. PLoS ONE 9: e Beal MS, Lattanzio MS, Miles DB Differences in the thermal physiology of adult Yarrow s spiny lizards (Sceloporus jarrovii) in relation to sex and body size. Ecology and Evolution 4: Bell RC, Zamudio KR Sexual dichromatism in frogs: natural selection, sexual selection and unexpected diversity. Proceedings of the Royal Society B: Biological Sciences 279: Berg W, Theisinger O, Dausmann KH Evaluation of skin temperature measurements as suitable surrogates of body temperature in lizards under field conditions. Herpetological Review 46: Brodie ED III, Janzen FJ Experimental studies of Coral Snake mimicry: generalized avoidance of ringed snake patterns by free-ranging avian predators. Functional Ecology 9: COLOUR CHANGE IN FENCE LIZARDS 407 Calisi RM, Hews DK Steroid correlates of multiple color traits in the spiny lizard, Sceloporus pyrocephalus. Journal of Comparative Physiology B 177: Carretero M Measuring body temperatures in small lacertids: infrared vs. contact thermometers. Basic and Applied Herpetology 26: Conant R, Collins JT A field guide to reptiles and amphibians of eastern and central North America, 3rd edn. New York: Houghton Mifflin. Cooper WE Jr, Burns N Social significance of ventrolateral coloration in the fence lizard, Sceloporus undulatus. Animal Behaviour 35: Cooper WE Jr, Greenberg N Reptilian coloration and behavior. In: Gans C, Crews D, eds. Biology of the reptilia, Vol. 18. Physiology E: hormones, brain and behavior. Chicago: University of Chicago Press, Crowley SR The effect of dessication upon the preferred body temperature and activity level of the lizard Sceloporus undulatus. Copeia 1987: Cuthill IC, Bennett ATD, Partridge JC, Maier EJ Plumage reflectance and the objective assessment of avian sexual dichromatism. American Naturalist 160: Doherty JA Temperature coupling and trade-off phenomena in the acoustic communication system of the cricket, Gryllus bimaculatus De Geer (Gryllidae). Journal of Experimental Biology 114: Endler JA On the measurement and classification of colour in studies of animal colour patterns. Biological Journal of the Linnean Society 41: Field A Discovering statistics using SPSS statistics (and sex and drugs and rock n roll), 3rd edn. London: Sage Publications. Geen MR, Johnston GR Coloration affects heating and cooling in three color morphs of the Australian bluetongue lizard, Tiliqua scincoides. Journal of Thermal Biology 43: Gerhardt HC Temperature coupling in the vocal communication system of the gray tree frog, Hyla versicolor. Science 199: Gilbert AL, Lattanzio MS Sceloporus undulatus (Eastern Fence Lizard). Novel ventral ornamentation. Herpetological Review 46: Gillooly JF, Ophir AG The energetic basis of acoustic communication. Proceedings of the Royal Society of London B 277: Grant BW Trade-offs in activity time and physiological performance for thermoregulating desert lizards, Sceloporus merriami. Ecology 71: Greenberg N Ethological aspects of stress in a model lizard, Anolis carolinensis. Integrative and Comparative Biology 42: Grether GF Carotenoid limitation and mate preference evolution: a test of the indicator hypothesis in guppies (Poecilia reticulata). Evolution 54: Haenel GJ, Smith LC, John-Alder HB Homerange analysis in Sceloporus undulatus. II. A test of spatial relationships and reproductive success. Copeia 2003:

13 408 B. P. STEPHENSON ET AL. Hare JR, Whitworth E, Cree A Correct orientation of a hand-held infrared thermometer is important for accurate measurement of body temperatures in small lizards and tuatara. Herpetological Review 38: Hews DK, Quinn VS Endocrinology of species differences in sexually dimorphic signals and aggression: using the organization and activation model in a phylogenetic framework. In: Fox SF, Baird TA, McCoy JK, eds. Lizard social behavior. Baltimore: Johns Hopkins University Press, Hill GE Female house finches prefer colourful males: sexual selection for a condition-dependent trait. Animal Behaviour 40: Hunt S, Cuthill IC, Bennett ATD, Griffiths R Preferences for ultraviolet partners in the blue tit. Animal Behaviour 58: Hutchinson VH, Larimer JL Reflectivity of the integuments of some lizards from different habitats. Ecology 41: Jensen JB, Camp CD, Gibbons W, Elliott MJ Amphibians and reptiles of georgia. Athens, GA: Georgia University Press. Jenssen TA, Greenberg N, Hovde KA Behavioral profile of free-ranging male lizards, Anolis carolinensis, across breeding and post-breeding seasons. Herpetological Monographs 9: Kour EL, Hutchison VH Critical thermal tolerances and heating and cooling rates of lizards from diverse habitats. Copeia 1970: Lakens D Calculating and reporting effect sizes to facilitate cumulative science: a practical primer for t-tests and ANOVAs. Frontiers in Physiology 4: 863. Langkilde T, Boronow KE Color as a signal: the relationship between coloration and morphology in male eastern fence lizards, Sceloporus undulatus. Journal of Herpetology 44: Langkilde T, Boronow KE Hot boys are blue: temperature-dependent color change in male eastern fence lizards. Journal of Herpetology 46: Langridge KV, Broom M, Osorio D Selective signalling by cuttlefish to predators. Current Biology 17: LeBas NR, Marshall NJ The role of colour in signalling and male choice in the agamid lizard Ctenophorus ornatus. Proceedings of the Royal Society of London B 267: Loew ER, Fleishman LJ, Foster RG, Provencio I Visual pigments and oil droplets in diurnal lizards: a comparative study of Caribbean anoles. Journal of Experimental Biology 205: Marler CA, Walsberg G, White ML, Moore M Increased energy expenditure due to increased territorial defense in male lizards after phenotypic manipulation. Behavioral Ecology and Sociobiology 37: Marsh R, Bennett A Thermal dependence of sprint performance of the lizard Sceloporus occidentalis. Journal of Experimental Biology 126: Mautz WJ, Daniels CB, Bennett AF Thermal dependence of locomotion and aggression in a Xantusiid lizard. Herpetologica 48: Metcalfe NB, Taylor AC, Thorpe JE Metabolic rate, social status and life-history strategies in Atlantic salmon. Animal Behaviour 49: Montgomerie R Analyzing colors. In: Hill GE, McGraw KJ, eds. Bird coloration, Vol. I: mechanisms and measurements. Cambridge, MA: Harvard University Press, Morrison RL, Rand MS, Frost-Mason SK Cellular basis of color differences in three morphs of the lizard Sceloporus undulatus erythrocheilus. Copeia 1995: Morrison RL, Sherbrooke WC, Frost-Mason SK Temperature-sensitive, physiologically active iridophores in the lizard Urosaurus ornatus: an ultrastructural analysis of color change. Copeia 1996: Norris KS Color adaptation in desert reptiles and its thermal relationships. In: Milstead WW, ed. Lizard ecology: a symposium. Columbia, MO: University of Missouri Press, O Connor K, Metcalfe N, Taylor A Does darkening signal submission in territorial contests between juvenile Atlantic salmon, Salmo salar? Animal Behaviour 58: Olson VA, Owens IPF Costly sexual signals: are carotenoids rare, risky or required? Trends in Ecology & Evolution 13: Olsson M, Stuart-Fox D, Ballena C Genetics and evolution of colour patterns in reptiles. Seminars in Cell & Developmental Biology 24: Plavicki J, Yang EJ, Wilczynski W Dominance status predicts response to nonsocial forced movement stress in the green anole lizard (Anolis carolinensis). Physiology and Behavior 80: Quinn VS Proximate and ultimate factors contributing to the variation in a sexually dimorphic signal in Sceloporus lizards. Unpublished D. Phil. Thesis, Indiana State University. Quinn VS, Hews DK Signal and behavioral response to the signal are not evolutionarily coupled in males: aggression affected by replacement of an evolutionarily lost colour signal. Proceedings of the Royal Society of London B 267: Rand MS Courtship and aggressive behavior in male lizards exhibiting two different sexual colorations. American Zoologist 28: 153A. Ressel S Sexual color dimorphism, sexual selection, and malarial parasitism in the western fence lizard, Sceloporus undulatus. Unpublished M. Sci. Thesis: University of Vermont. Ressel S, Schall JJ Parasites and showy males: malarial infection and color variation in fence lizards. Oecologia 78: Rice GE, Bradshaw SD Changes in dermal reflectance and vascularity and their effects on thermoregulation in Amphibolurus nuchalis (Reptilia: Agamidae). Journal of Comparative Physiology 135: Ries C, Spaethe J, Sztatecsny M, Strondl C, H odl W Turning blue and ultraviolet: sex-specific colour change during the mating season in the Balkan moor frog. Journal of Zoology 276: