Effects of substrate color on intraspecific body color variation in the toad‐headed lizard, Phrynocephalus versicolor

Abstract Diversity in animal coloration is generally associated with adaptation to their living habitats, ranging from territorial display and sexual selection to predation or predation avoidance, and thermoregulation. However, the mechanism underlying color variation in toad‐headed Phrynocephalus lizards remains poorly understood. In this study, we investigated the population color variation of Phrynocephalus versicolor. We found that lizards distributed in dark substrate have darker dorsal coloration (melanic lizards) than populations living in light substrates. This characteristic may improve their camouflage effectiveness. A reciprocal substrate translocation experiment was conducted to clarify the potential role of morphological adaptation and physiological plasticity of this variation. Spectrometry technology and digital photography were used to quantify the color variation of the above‐mentioned melanic and nonmelanic P. versicolor populations and their native substrate. Additionally, substrate color preference in both populations was investigated with choice experiments. Our results indicate that the melanic and nonmelanic populations with remarkable habitat color difference were significantly different on measured reflectance, luminance, and RGB values. Twenty‐four hours, 30 days, and 60 days of substrate translocation treatment had little effects on dorsal color change. We also found that melanic lizards choose to live in dark substrate, while nonmelanic lizards have no preference for substrate color. In conclusion, our results support that the dorsal coloration of P. versicolor, associated with substrate color, is likely a morphological adaptation rather than phenotypic plasticity.

Camouflage is a key for animals to survive in different color backgrounds. To display a feature of camouflage, it is of great importance for animals to select a substrate that blends with their body coloration.
The skin of reptiles is generally considered to have three layers of pigment-containing and light-reflecting cells (chromatophores). The melanophore is in the third layer of skin, which is responsible for the synthesis of melanin (Rosenblum, Hoekstra, & Nachman, 2004). Two principal types of animal color changes have been reported: physiological and morphological color change. Physiological color change is caused by the movement (aggregation or dispersion) of pigment granules within chromatophores, often taking place in milliseconds to hours in response to short-term environmental stimuli, showing a strong plasticity on animals' body color (Nery & Castrucci, 1997;Sköld, Aspengren, Cheney, & Wallin, 2016;Sköld, Aspengren, & Wallin, 2013). In contrast, morphological color change occurs due to changes in the pigment production of chromatophores and takes up to several days or months in response to long-term environmental stimuli. It is generally considered to be long-term adaptation to a certain background (Kraemer, Kissner, & Adams, 2012;Rosenblum, 2005). Since the short-term plasticity or long-term adaptation would, respectively, result in physiological or morphological color change, the type of change within a specific organism needs to be clarified to point out the direction for future research.
The Phrynocephalus lizards have received ecological attention on color variation in recent years. The central black abdomen existing only in high altitude populations of P. theobaldi was discovered by Jin and Liao (2015), and their subsequent study further confirms that this trait is associated with thermoregulation in cold regions . Sexual color dimorphism was reported in P. guinanensis (Ji, Wang, & Wang, 2009;Zhang, Tong, Wo, Liu, & Jin, 2018).
Moreover, Tong et al. (2016) described the dorsal gray color variation between P. versicolor and P. frontalis. However, little is known about intraspecific color variation associated with habitat color changes in this genus. Phrynocephalus versicolor widely inhabits the deserts and semideserts endemic in eastern Xinjiang, western Inner Mongolia, western Gansu, and Ningxia province in China (Zhao, Zhao, & Zhou, 1999). We observed the dorsal coloration of P. versicolor distributed only at dark substrate in Liuyuan town (Gansu province) is visibly darker (melanic, see Figure 1a) than other conspecific populations (nonmelanic, see Figure 1a) living in weathered yellow (light) substrate. In this study, spectrometry technology (Matthews, Goulet, Delhey, & Chapple, 2016;Rowe, Bunce, & Clark, 2014;Rowe, Miller, et al., 2014) and digital photography (McGauch, 2008;McKay, 2013;Stevens, Parraga, Cuthill, Partridge, & Troscianko, 2007) were used to quantify the natural dorsal color variation and their native substrate color difference between melanic and nonmelanic P. versicolor populations, aiming to determine: (a) whether the color properties of melanic and nonmelanic lizards differed significantly; (b) whether the dorsal color properties from different habitats were associated with the different color substrates (i.e., black and weathered yellow sand), which could suggest background matching camouflage; and (c) whether each lizard color morph displays a behavioral preference for substrates that match their body coloration.

| Sampling
In May, 2017, our first sampling trip, a total of 45 (18 males, 27 females) melanic P. versicolor adults were collected from black mountainous In July, 2018, our second sampling trip, a total of 20 nonmelanic P. versicolor adults were collected from Guazhou county, Gansu, China (GZ, 95.61°E, 41.05°N, 1,386 m a.s.l., 11 males, 9 females) and 24 melanic P. versicolor adults (6 males, 18 females) from HSK. All adult lizards were collected within four days of fieldwork for each trip.
The melanic population was only observed in HSK, where the substrate is black. We found three nonmelanic populations had similar dorsal color (nonmelanic) and substrate color (light), despite the large varied elevation and geographical distance that separates them ( Figure 1). All the collected individuals were transported and fed in the zoology laboratory of Lanzhou University within 24 hr after captured, and all lizards used in this work were individually marked (toe-clipped) to give their ID. After the experiments were performed, they were safely released back into the wild at their captured places.

| Feeding and reciprocal translocation experiments in artificial laboratory conditions
We used rectangular light plastic boxes (0.79 m × 0.60 m × 0.50 m) to house captured lizards. Each box was covered by 3-to 5-cm-thick sand and stones, that is, using black versus weathered yellow stones collected from natural habitats of populations to, respectively, simulate natural dark versus light habitat types ( Figure 2a). Partial regions of each box were in direct sunlight for 1-2 hr during the day to allow lizards' body temperature thermoregulation. Lizards were fed daily with Tenebrio molitor larvae and had a permanent supply of water. Prior to experiments, all melanic and nonmelanic P. versicolor adults were regularly fed in black and weathered yellow habitat, respectively. During the translocation experiments, melanic population was housed in a weathered yellow sand habitat, while the nonmelanic populations were kept in black substrate. Each box housed eight to nine individuals. Males and females of the same population were allocated evenly in independent boxes. A total of 45 melanic P. versicolor adults (HSK population, 18 males, 27 females) and 78 nonmelanic adults (EJN: 11 males, 23 females; SS: 11 males, 33 females) from the first sampling trip were fed over one month. In addition, 24 melanic adults (HSK population, 6 males, 18 females) from the second sampling trip were raised over two months.

| Dorsal color measurements
Comparisons of dorsal color between melanic and nonmelanic populations were conducted independently through two sets of measurements. The first set was to measure reflectance of HSK and EJN/SS populations using the spectrophotometer. The second was to analyze RGB, and luminance values estimated from photographs of other individuals of HSK and GZ populations using Adobe Photoshop SC6.
RGB values provide color information which could show additional variation in hue compared with dark lizards. The former method using fiber spectrophotometer is very sensitive for color changes and easy to finish white color calibration. Moreover, the reflective probe should be very close to the measuring object in the presence of alternating current during measurement. It is thus good to be used in laboratory environment. However, it has disadvantages to obtain coloration data on the whole body conveniently. In contrast, the latter method has the obvious advantage to measure dorsal and substrate color immediately and conveniently during field work.

| Photographs of substrate and lizards in field
To compare the dorsal color variation between black and weathered yellow substrates, a total of 20 and 24 digital photographs from the HSK and the GZ areas, respectively, were taken. Since no significant color variation existed among different nonmelanic populations (GZ, SS, and EJN) or their substrates (Figures 1a and 2b), here we only used GZ population which was geographically closest to the HSK lizards. Each digital photograph represents the place of substrate where lizards were discovered. Another 10 digital photographs with each of them containing both a lizard and the substrate where it was captured were taken from each of HSK and GZ to analyze the relationship between dorsal color and substrate color. All these photographs were taken from 9:00 a.m. to 11:00 a.m. within one day of fieldwork in July 2018. A black umbrella was used to block the direct sunlight. All these photographs were taken by the same camera with identical photographic parameters and were analyzed with the same color calibration described above.

| Substrate selection
To investigate whether melanic HSK and nonmelanic GZ P. versicolor adults were behaviorally segregated based on substrate color preference, according to their dorsal coloration, a choice experiment was performed to allow lizards choose between dark or light substrate.

| Data analysis
The spectroscopic data, which included reflectance values with wavelength ranging from 300 to 700 nm, were extracted and

| Color variation of substrates and lizards
Dorsal and substrate color were compared between the geographically nearby melanic HSK population and the nonmelanic GZ popula-

| Reflectance variation between melanic and nonmelanic populations
Before the treatment of substrate translocation, no significant differences of reflectance between male and female (melanic:

| Effects of habitat color on body color variation
For HSK population, no significant reflectance differences were found by single linear mixed model analysis between 0 hr and 24th

| Substrate color selection
Melanic lizards tended to prefer the black rather than weathered yellow substrate (t 19 = 6.200; p < .001), while no significant preference of substrate color was detected for nonmelanic GZ lizards (t 14 = 1.025; p = .323; Figure 3b), based on paired t tests.

| D ISCUSS I ON
Using spectrometry technology and digital photography, we analyzed for the first time the detailed body coloration and behavior responses of different P. versicolor populations to substrate color. We found that the body coloration of P. versicolor from dark substrate was conspicuously melanic compared with nonmelanic conspecifics from light substrate, whether they experienced the substrate translocation treatment or not. The lizards had little changes on the reflectance of their dorsal skin, and no macroscopic color changes were observed after translocation treatment of substrates. Also, melanic lizards prefer to live in black substrate, but nonmelanic lizards are free of bias for substrate color.
Variation of animal coloration has been described to be correlated with climatic factors, such as the description of Gloger rule, which states that more heavily pigmented forms within one species tend to be found in more humid environment. However, it is often considered to be associated with substrate color, that is, the lizard color variation in white sands (Rosenblum et al., 2010 Figure 1) would most likely be the primary factor, though temperature (Krohn & Rosenblum, 2016;Rosenblum, 2005) and stress (Krohn & Rosenblum, 2016) may affect lizards' body color. Nonmelanic populations from large areas with variable elevation-associated environmental conditions (Figure 1b) show similar light dorsal color, while the melanic population in HSK exhibits dark dorsal color in high altitude as well (Figure 1a), indicating that elevation is probably not contributing to explain the variance of coloration. Moreover, substrate color difference has been previously found to explain coloration variation within a species between geographically nearby populations, for example, Anolis carolinensis (Macedonia, Echternacht, & Walguarnery, 2003), Crotalus lepidus lepidus (Farallo & Forstner, 2012), and Montivipera raddei species complex (Rajabizadeh et al., 2015). We propose that this is the same case for melanic HSK and nonmelanic GZ populations.
Surprisingly, the melanic P. versicolor did not increase their reflectance value after 24 hr of treatment in weathered yellow substrate, and the same results were obtained for nonmelanic populations which did not decrease their reflectance value after translocation to dark substrate. Moreover, the melanic and nonmelanic populations still, respectively, kept dark and light dorsal color after 60 or 30 days of translocation treatment of substrates (Figure 2), showing slight phenotypic plasticity on dorsal color could occur in this case. This finding support that dorsal color variation in P. versicolor is inclined to morphological adaptation rather than physiological plasticity.
However, it is worthy to note that several measuring points showed significantly statistical differences between pretreatment and posttreatment, and we therefore speculate that it may be induced by a certain degree of aggregation and dispersion of melanin granules in these body parts. As this work is primarily focused on ecological expectations of body coloration in P. versicolor, this scenario could be explored in the future with tissue experiments, where the number of chromatophores in corresponding parts of tissues is measured.
The melanic degree of animal skin is often associated with the activity of melanophores (Alibardi, 2013), and it may take several days or even month for epidermis melanophore layer to divert melanosome to adjacent keratinocyte layer, rendering the skin melanism, that is, morphological color change (Cooper & Greenberg, 1992).
Selective stress may influence the regulation of hormones and other potential genetic factors that are important for pigment synthesis for example, α-MSH (Fernandez & Bagnara, 1991;Mashinini, Heideman, & Mouton, 2008). In several classic studies, individual body coloration would become similar with local substrate environment when they subject to stronger pressure from predators (Cooper & Allen, 1994;Johnsson & Kjällman-Eriksson, 2008;Kettlewell, 1973).
Therefore, this could be the case for HSK P. versicolor populations since no melanic individuals were found outside of this area, whose dark body color tends to be influenced by evolutionary factors caused by long-term substrate color stimuli, rather than the physiological changes in response to short-term substrate color changes.
Further research is needed to investigate the predator pressure and the underlying genetic factors that might induce the geographical dorsal color variation of P. versicolor populations.
In animals, behaviorally mediated background matching is a crit-  Jin et al., 2016), for example, the functions of abdominal black-speckled area reported in P. theobaldi (Jin & Liao, 2015). However, almost all melanic and nonmelanic P. versicolor populations only survive at dark and light substrates, respectively, and moreover, nonmelanic P. versicolor populations did not display obvious color differences in large different elevation/temperature environments. Therefore, the primary pressure for melanic lizards is black substrate rather than environmental temperature.
Such type of body coloration, combined with their native substrate color and their behavior for background matching, could improve the effectiveness of successful camouflage. Notably, nonmelanic P. versicolor lizards did not seem to have preference for weathered yellow substrate, as the lizard frequency in both types of substrates is similar (Figure 3). These results may indicate that the behaviorally mediated substrate preference of melanic P. versicolor population may evolve from phenotypic plasticity on dorsal coloration in their initial colonization, similar with the dorsal plasticity of a population of side-blotched lizard, Uta stansburiana (Corl et al., 2018).
In conclusion, we discovered that substrate color associated dorsal coloration of P. versicolor is most likely a morphological longterm adaptation in response to heterogeneous substrate color, which could help their camouflage effectiveness. Our study firstly addressed the natural dorsal color variation among intraspecific populations of Phrynocephalus. Our findings provide insights into our understanding on why pigmentation of Phrynocephalus in dark substrate emerged and may be useful for further identification of potential genetic factors responsible for morphological adaptation in a genomic and functional way.

ACK N OWLED G M ENTS
This work was supported by the National Natural Science Foundation of China (31772447). We wish to thank Kai-Long Zhang, Shao-Hua Yan, and Yu-Hang Liu for help in specimen collecting and laboratory work with the corresponding author.

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
HJT, JSL, YBW, GS, and YTJ contributed experiments, analyses, and materials; HJT, AGD, and YTJ wrote and revised the manuscript; YTJ designed the study; and WZ contributed materials.