An elevational trend of body size variation in a cold-climate agamid lizard, Phrynocephalus theobaldi

UNCORRECTED PROOF An elevational trend of body size variation in a coldclimate agamid lizard, Phrynocephalus theobaldi Yuanting JIN, Pinghu LIAO College of Life Sciences, China Jiliang University, Hangzhou, 10018, China Abstract The pattern that many ectotherms have smaller body sizes in cold environments follows the converse to Bergmann s rule and is most frequently found in lizards. Allen s rule predicts animals from warm climates usually have longer tails and limbs, while these traits tend to be shorter in individuals from cold climates. We examined body size variation in an endemic Chinese lizard, Phrynocephalus theobaldi, along a broad elevational gradient (600 5000 m on the QinghaiTibetan Plateau). Female body size showed a Ushaped cline, decreasing with increased elevation within the range 6004200 m, but increasing at elevations > 4200 m. Male body size continued to increase with increasing elevations. Both sexes showed an increased pattern of extremity length with elevation that does not conform to Allen s rule. Limb length and tail length increased along the elevational gradients. In terms of color pattern, an abdominal black speckled area appears at elevations >4200 m. This trait increases in size with increased elevation. Unlike most studies, our results indicated that annual sunshine hours corresponding to the activity period of the lizards could play an important role on the positive body size cline in environments at very high elevations > 4200 m [Current Zoology 61 (2) :, 2015 ]. Keywords Allen s rule, Bergmann s rule, Elevation, Thermal constraint, Tibetan Plateau, Toadheaded Agama Understanding body size variation in a spatial context can offer key insights into the process of evolution. Bergmann s Rule attempts to explain variation in body size while Allen s Rule aims to explain body shape, but their applicability to ectotherms is debatable. The former states that animals living in cold climates tend to be larger than individuals of the same species living in warm climates (Bergmann, 1847). Among reptiles, some squamates tend to follow Bergmann s Rule (Angilletta et al., 2004; Du et al., 2005; Roitberg et al., 201; Feldman and Meiri, 2014; Received June 2, 2014; accepted Aug. 20, 2014. Corresponding author. Email: jinyuanting@126.com Current Zoology 2015 0

Volynchik, 2014), while others appear to show the opposite pattern (Ashton and Feldman, 200; PincheiraDonoso et al., 2008; Brandt and Navas, 201; PincheiraDonoso and Meiri, 201). Although patterns of body size variation and their relationships with environmental factors have been studied in ectothermic vertebrates (Nevo, 1981; Dunham et al., 1989; Atkinson, 1994; Beaupre, 1995; Angilletta and Dunham, 200; Stillwell and Fox, 2009; Roitberg et al., 201), most studies have focused on latitudebased differences (Angilletta et al., 2004; Sears and Angilletta, 2004), and some researches showed sexspecific phenotypic response to temperature (i.e., Volynchik, 2014). By comparison, little research has been conducted along elevational gradients (but see, PincheiraDonoso et al., 2008; Hu et al., 2011). Allen s Rule states that endotherms from colder climates usually have shorter limbs (or appendages) than the equivalent animals from warmer climates (Allen, 1877). The explanation of the rule is that endothermic animals with the same volume may have different surface areas, which will aid or impede their temperature regulation. Ectotherms with different surface area to volume ratios will exchange heat at different rates, which could create different selection pressures in different thermal environments. In lizards, the relative extremity lengths of lacertids, especially limb segments, has been observed to generally increase towards hot locations, following Allen s rule (Volynchik, 2014). Research on the body sizes of P. vlanglaii has documented relatively short arm and leg lengths in warm environments, contrary to the predictions of Allen s rule (Jin et al., 2007). Variation in body size of ectotherms along latitudinal and elevational clines is assumed to represent the adaptation of populations to local environmental conditions (Ashton and Feldman, 200). Temperature is frequently cited to explain latitudinal variation in snoutvent length (SVL) for reptiles (Du et al., 2005; Volynchik, 2014). Factors such as food availability associated with different activity periods or reproductive states are also invoked to explain intraspecific size clines in the direction predicted by Bergmann s rule (Ashton and Feldman, 200; PincheiraDonoso and Meiri, 201). The environmental features that generate geographic variation in body size, however, might differ considerably between elevational and latitudinal ranges. For example, variation in hypoxia, sunshine, rainfall and insolation at high elevations are not always replicated at high latitudes (Körner, 2007). Known as the "roof of the world", the QinghaiTibetan Plateau (QTP) in southwestern China is the highest and largest plateau on earth (Tapponnier et al., 2001). The flat toadheaded lizard genus Phrynocephalus is the most abundant and has the widest distribution of all reptiles found on the QTP (Zhao et al., 1999). Six morphologically defined viviparous species (P. vlangalii, P. erythrurus, P. theobaldi, P. putjatia, P. forsythii and P. guinanensis) are present on the QTP and form a monophyletic clade within the genus (see Jin and Brown, 201). Phrynocephalus viviparous lizards often exhibit blackspeckled areas along the abdomen, predicted to be important adaptations when basking in cold environments (Zhao et al., 1999). To date, the morphological and reproductive characteristics of only a 1

single species, P. vlangalii, has been intensively studied (Huang and Liu, 2002; Wu et al., 2002, 2004; Zhang et al., 2005). Intraspecific variation in body size of P. vlangalii does not appear to conform to Bergmann s Rule: analyses of two intraspecific lineages, P. vlangalii vlangalii and P. vlangalii pylzowi that occupy elevations < 4250 m show similar but nonsignificant negative relationships between SVL and elevation which runs counter to the pattern expected under Bergmann s rule (Jin et al., 201). Further research on this genus is needed, however, to elucidate whether this pattern occurs generally. The toadheaded agama, P. theobaldi is viviparous and distributed in the western and southern portion of the Xizang area of the QTP, including Ngari areas in western Tibet, the Brahmaputra River valley, and the Xizang Southern Valley, between the north side of the Himalaya mountains and the south side of the GangdiseNyainqentanglha mountains (Zhao et al., 1999). This species is diurnal and occurs across small patches of sand dune habitats, semideserts or rocky steppe habitats (Zhao et al., 1999). It has two subspecies, P. theobaldi theobaldi (ranging across higher regions of Ngari in the western Xizang area) and P. theobaldi orientalis (occuring in the lower regions of Brahmaputra River valley and the Xizang Southern Valley; Wang et al., 1999). The morphological differences between the two subspecies are: the tips of the hind limbs in P. theobaldi orientalis do not reach the mouth, but touch the mouth margin in P. theobaldi theobaldi; and the length of the tail of P. theobaldi orientalis is shorter than its SVL, whereas it exceeds SVL in P. theobaldi theobaldi (Wang et al., 1999). An implication from these descriptive traits is that populations of P. theobaldi could show size variation along a very high and broad elevational gradient (600 5100 m). Whether or not the species shows decreasing extremity length with elevation has not yet been examined. Here, we investigated intraspecific size variation in the species along elevational gradients and examined the potential influence of environmental climatic factors on size evolution. The specific objectives of this work were to: (1) assess whether the body size variation of P. theobaldi contradicts that predicted by Bergmann s Rule (as is often found in lizards); (2) examine whether or not the species conforms to the prediction of Allen s rule that high elevation populations have relatively shorter limbs or tails; and () verify the potential influence of climatic factors along elevational gradients on the body size cline. 1 Materials and Methods Specimens were collected along the entire elevational gradient (,600 5,050 m) of the range of P. theobaldi in western and southern QTP, during the breeding period in August 2011. Elevations of sampled sites were recorded using a Garmin Oregon 400t handheld GPS unit. Sample sites were grouped into 19 geographical sampling areas, with each area covering an elevational range 50 m, and an area 5 ha. This delineation was chosen because we could not locate any populations having individuals spread over greater elevational gradients due to habitat isolation. 2

We analyzed overall body size variation along elevational gradients by measuring the following morphological traits (± 0.01 mm): snoutvent length (SVL), tail length (TL, from vent to the tail tip), head length (HL, from tip of snout to posterior margin of occipital scale), head width (HW, taken at the posterior end of the mandible), arm length (AL, distance between axilla and wrist), leg length (LL, distance between groin and ankle), distance between axillae (DBA) and distance between iliac crests (DBI). All traits were measured by the same person using ShangliangVernier Calipers on living specimens. The measurements were checked and confirmed to be consistent across individuals by the same author with help from another. Neonate, juvenile and adult individuals are distinguishable from each other in the field based on body size. Lizards that were longer (in SVL) than the smallest gravid females of the lizard were considered to be adults (Zhang et al., 2005). Individuals usually become sexually mature at about two years of age (Zhao et al., 1999). Body size analyses were based on adults only. Climate data over 0 years from 27 climatic stations in the Xizang portion of the QTP were collected from the Chinese National Climatic Data Center (CDC). Of the 27 stations, 20 of them occurred within the potential distribution range of this species. In order to survey the climatic factors of our sampling areas and to analyze their relationships with body size of the lizards, we calculated monthly means of the following climatic variables which could potentially influence the body size of animals: temperature ( C), barometric pressure (hpa), rainfall (mm), relative humidity (%), and hours of sunshine. We calculated the 0year mean values of these variables for each month, from May to September (corresponding to the activity period of the lizards). We analyzed relationships between elevations of climatic stations and mean climate values using bivariate correlation and partial correlation, with or without control for latitude or longitude. Among these climatic factors, we found that only temperature and barometric pressure significantly decreased with increasing elevation, and hours of sunshine increased significantly with increasing elevation. Thus, only these climatic factors were included because of their relevance to the aims of this study. We also regressed these factors on elevation to obtain regression equations that would allow estimation of the climate in sampling areas lacking climate data. Climatic values for P. theobaldi sampling areas were obtained from regressions of the factors on elevation. The linear regressions can be summarized as: Temperature ( C) =29.89080.0049ELE, r 2 = 0.78, F 1,25 =86.42, P< 0.001; Barometric Pressure (hpa) = 945.1870.0792ELE, r 2 = 0.99, F 1,24 =626.4, P< 0.001; hour of sunshine (h) = 255.882+0.050ELE, r 2 = 0.72, F 1,25 =14.58, P< 0.001. In our study area, rainfall increased with increasing longitude (partial correlation: r=0.754, P<0.001) but had a nonsignificant relationship with elevation (partial correlation: r=0.01. P=0.880) when longitude or altitude was controlled for. Analyses were conducted with SPSS v19.0.

SVLs of both sexes were compared using oneway analyses of variance (ANOVA). There was a significant interaction between sex and elevation (F 15, 15 =1.846, P=0.028), indicating different male and female responses to elevation. SVL and other characteristics of each sex were used in subsequent analyses. We used SVL as a surrogate for body size, but considered principle component analysis (PCA) to be a more effective method for examining generalized lizard size (McCoyet al., 2006). Principle components (PCs) for each sex were obtained by using the correlation matrix of the eight morphological variables. Only PCs with eigenvalues greater than 1.0 were used for subsequent analyses. High loading values for all morphological traits on PC1 indicated that it was a suitable surrogate for overall size in our study. Scatterplots indicated that the relationship of female is more curvilinear than linear and indicated Ushaped patterns. Clinal patterns of SVL and PC1 scores along elevational gradients were tested using correlation analyses. A total of 11 models were tested using linear and nonlinear regression, and the corresponding AIC value of each model was calculated. All correlation or regression analyses were performed using mean values (all reported ±1 SD). Probability values < 0.05 were considered to be statistically significant. The individual abdominal blackspeckled area (ABA) was approximately rectangular and so its size was calculated as length multiplied by width. The relationship between the mean abdominal blackspeckled area and elevation was analyzed using partial correlation analysis while controlling for SVL. 2 Results 2.1 Body size and sexual size dimorphism Mean values for SVL across all sampling areas ranged from 46.0 52.04 mm (Table 1). Mature female SVL ranged from 42.0 mm to 6.4 mm while mature male SVL ranged from 42. to 58.7 mm. Values for both female (49.2 ± 4.1 mm, n = 175) and male (49.1 ±.8 mm, n = 175) subjects were similar (oneway ANOVA: F 1, 48 = 0.146, P = 0.70). Female and Male PC1 had similarly high loading values for all variables, indicating that they represented generalized size (Table 2). The amount of variation in body size was similar for both sexes of lizards (Table ). 2.2 Body size variation with elevation and climatic factors In general, correlation analyses showed positive relationships between elevation and SVL and PC1. Both SVL and PC1 showed similar clinal patterns with elevation in both sexes (Fig. 1). A linear regression model showed positive trends on the elevational clines of SVL and PC1 (Table 4). A curvilinear model appears to provide the best fit for females (Fig. 1), specifically the significant quadratic and cubic models had the smallest AIC values of the 11 models tested. Adding quadratic and cubic terms to the regressions also resulted in a greater proportion of the variation being explained (Table 4), showing a U shaped cline in female body size along elevational gradients. The curve inflexion 4

point for the regression of SVL and PC1 on elevation occurs at about 4200 m (Fig. 1). For the portion of the curve spanning 600 4250 m in elevation, both female SVL and PC1 showed nonsignificant negative correlations with elevation (Pearson correlation SVL: r=0.027, P = 0.960, n=6; PC1: r=0.108, P = 0.88, n=6). Between 42505050 m, both SVL (r=0.778, P=0.00, n=12) and PC1 (r=0.740, P=0.006, n=12) increased with increasing elevation. Male SVL and PC1 increased with increasing elevation (Pearson correlation SVL: r=0.584, P = 0.014, n=17; PC1: r=0.4, P = 0.177, n=17). These results indicate that body size generally increases with increased elevation, for both sexes. Male SVL and PC1, together with female (> 4200 m) SVL and PC1 decreased with increasing temperature and barometric pressure, and increased with increasing hours of sunshine. There is no significant relationship between female (< 4200 m) SVL and PC1 and these three climatic factors (Table 5). Partial correlation analyses in which the correlation with SVL was removed showed significant relationships between some traits and elevation (For female TL: r = 0.51, P = 0.05; HL: r = 0.76, P < 0.001; AL: r = 0.50, P = 0.029; LL: r = 0.680, P = 0.00; DBA: r = 0.604, P = 0.010; and for Male TL: r = 0.48, P = 0.05; AL: r = 0.801, P < 0.001; LL: r = 0.748, P < 0.001; DBI: r = 0.511, P = 0.04). These analyses supported increasing TL, AL and LL with increasing elevation for both sexes. 2. Variaton of adbominal blackspeckled area with elevation The mean abdominal blackspeckled area increased with increasing elevation when mean values for SVL were controlled in the analyses (partial correlation: r=0.725, P<0.001 for female; r = 0.642, P < 0.001 for male). This correlation was also significant when populations without blackspeckled areas, i.e., populations < 4,200 m were included in the analysis (this involved the inclusion of many individuals with zero scores for this trait). Discussion We found a Ushaped relationship between body size and elevation in P. theobaldi: large female body size was found at the lowest and highest gradients of very high elevations and small female body size at intermediate elevations. To our knowledge, such a pattern has not been observed before with regard to body size and elevation in lizards. Some studies have detected linear, not Ushaped curve relationships (Jin et al., 2007; PincheiraDonoso et al., 2007). Our results showed that the species cline provides only partial support for Bergmann s rule. A number of hypotheses have been proposed to explain the influence of environmental factors on animal body size. Bergmann s rule predicts that body size will be positively correlated with latitude and elevation and is based on the premise of heatconservation. An organism s thermal environment differs from some other sources of selection in that it is a dominant selective force at low temperatures (Ashton and Feldman, 200; PincheiraDonoso et al., 2007). At higher temperatures, ectotherms had relatively little thermal constraints, and hence other factors may become dominant in 5

determining body size. Phrynocephalus theobaldi does not possess an abdominal black speckled patch at lower elevations (< 4,200 m) in the Tibetan Southern Valley, but this trait was present at elevations > 4200 m and increased in size at higher elevations. This could reflect weak thermal constraints on individuals occurring below 4,200 m, because the abdominal black area might help thermoregulation of Phrynocephalus (data unpublished, personal communication). Other climatic factors might play a more important role on this portion of cline. Hypoxia is another prominent stress on individuals living at high elevations (Hammond et al., 2001). The oxygen consumption of ectotherms is lower under hypoxic conditions (Van den Thillart et al., 1992). The same quantity of food consumed at higher elevations by reptiles will produce less energy than at normal elevations (Dawson, 1975; Grant, 1990). Moreover, along the cline in the Tibetan Southern Valley with elevations < 4200 m (having greater precipitation and fewer hours of sunshine compared to other areas of distribution with elevations >4,200 m in the QTP), the influence of these other factors could obscure the positive role of hours of sunshine on increased growth < 4,200 m. The decreasing body size cline observed in similar elevational gradients < 4,200 m for two lineages of P. vlangalii vlangalii (2,756,470m) and P. vlangalii pylzowi (2,926 4,250m; Jin et al., 201) appears to be consistent with the pattern observed in female SVL < 4200 m, although the significance of this patterns cannot be established due to a lack of statistical power. Phrynocephalus theobaldi shows clinal variation at elevations > 4,200 m which is the opposite to the pattern seen in most squamates (Ashton and Feldman, 200; Angilletta et al., 2004). Almost without exception, the traditional heat conservation explanation (Bergmann, 1847; Mayr, 1956) does not apply to squamates (Ashton and Feldman, 200; de Queiroz and Ashton, 2004; PincheiraDonosoet al., 2007). In fact, low temperatures and hypoxia at high latitude and/or elevation are not conducive to increased growth, because temperature is closely related to the activity, feeding and digestion ability, and also to net energy availability for growth (Dawson, 1975; Dunham et al., 1989; Grant and Dunham, 1990; Brown and Griffin, 2005). Although resting metabolic rate might be lower for lizards at lower temperatures, in colder environments, the net energy gained from the same food will drop, so this is not necessarily conducive to increased availability of resources for growth (Dawson, 1975; Grant, 1990; Grant and Dunham, 1990). The general consensus is that temperaturebased lower rates of seasonal activity for the lizards at high elevation or latitudes will negatively affect body size growth (Hellmich, 1951; Grant and Dunham, 1990; Sears and Angilletta, 2004). Moreover, a lizard s ability to capture prey, avoid predation, maintain endurance, attract mates, produce sperm, and convert lipids are substantially reduced at suboptimal body temperatures (Shine, 1980; Hailey and Davies, 1988; MartinVallejo et al., 1995; Pianka and Vitt, 200; Ibargüengoytía, 2005). However, part of our study does show a different pattern, where populations showed increased body sizes at the very highest elevations (>4,200 m). Although there is little available information in the literature (but, see Heath, 1965; Muth, 1977), the duration of 6

available sunshine could be a key environmental factor. In our study area with very high elevations, duration of sunshine period was positively correlated with elevation. Moreover, the abdominal blackspeckled patch present on lizards at > 4,200 m (and positively correlated with elevation) helps absorb the heat radiating from the ground surface. On clear days, the sands substrate temperature measured during field work was usually higher than the air temperature. In other words, individuals at relative higher elevations might have longer periods available for foraging activity, compared to those at relative lower elevations, because of a greater ability to thermoregulate with their abdominal patches and relative longer tails and limbs, implying larger surface area to volume ratios of P. theobaldi. On cloudy days, the highest mean air temperature during the activity period of the lizard was just 12 C. It is lower than the ground surface temperature when the species was frequently active (the temperature of the sand substrate often reached 0 C on sunny days). This implies that sunshine may play a dominant role on this cline > 4,200 m compared to the factors as temperature and hypoxia. Our data also contrast with the predictions of Allen's rule (e.g., shorter appendages at higher latitudes/altitudes). Longer appendages for the lizard may work to improve heat gain during basking activity. In cold environments, adaptations to heatgain are under strong positive selection to allow optimal feeding, mating and predator avoidance (PincheiraDonoso et al., 2008). Also, the longer tails and limbs in relative high elevations might compensate weakened motor ability in hypoxia environments. Rainfall and air humidity can influence animal body size at hatching (Janzen et al., 1995; Du and Shine, 2008), or not (Brana and Ji, 2000; Flatt et al., 2001). In our study area of the QTP, rainfall did not significantly correlate with elevation, indicating a lower influence on these clines. Other factors that have been previously shown to influence a body size cline (but which we did not measure) include maturation rate (Atkinson, 1994; Belk and Houston, 2002; Palkovacs, 200), phylogenetic constraints (de Queiroz and Ashton, 2004;Gaston and Blackburn, 2008), migration pattern (Gaston and Blackburn, 2008), resistance to starvation (Lindstedt and Boyce, 1985; Cushman et al., 199; Ashton and Feldman, 200), and effects on somatic cell sizes (Van Voorhies, 1996, 1997; Mousseau, 1997; Partridge and Coyne, 1997). Nevertheless, it is unusual for lizards to show the body size trends predicted by Bergmann's rule (Ashton, 2001; Espinoza et al., 2004; Cruz et al., 2005). The body size cline could be partially mediated by the thermal environment and hypoxia at sites < 4200 m, and by the duration of available sunshine at sites > 4,200 m. The historical relationships among P. theobaldi populations are not known. However, the phylogenetic effects on this body size cline are likely to be minor because of the small difference in body sizes between populations. Bergmann or Allen s rules have usually been tested in situations far less restricted than our study. It is not surprising that we have found a distinct pattern from other studies considering 7

Bergmann and Allen's rules, because our study is limited to very high elevations that are absent in most parts of the world. As sample sizes were relatively small (< 5 for each sex) at 10 of the 19 sampling sites, future work will involve increasing the statistical power of our analyses by adding more specimens. Acknowledgments This work was supported by the National Natural Science Foundation of China (17218, 1000950). We thank the Chinese National Climate Data Center for historical climatic data, and the Tibetan government for permission to collect specimens. We thank R. Brown, S. Mullin for their kind detailed reviews on the manuscript before submission, and appreciate R. Brown for his extensive help in improving the language of the manuscript before publication. We also thank S. Meiri and another anonymous referee for their kind revisions or suggestions on our submitted manuscript. References Allen JA, 1877. The influence of physical conditions in the genesis of species. Radical Review 1: 108 140. Feldman A, Meiri S, 2014. Australian snakes do not follow Bergmann s rule. Evolutionary Biology 41: 27 5. Angilletta MJ Jr, Dunham AE, 200. The temperaturesize rule in ectotherms: Simple evolutionary explanations may not be general. American Naturalist 162: 2 42. Angilletta Jr MJ, Niewiarowski PH, Dunham AE, Leaché AD, Porter WP, 2004. Bergmann s clines in ectotherms: Illustrating a lifehistory perspective with Sceloporine lizards. American Naturalist 164: E168 E18. Ashton KG, 2001. Body size variation among mainland populations of the western rattlesnake Crotalus viridis. Evolution 55: 252 25. Ashton KG, Feldman CR. 200. Bergmann's rule in nonavian reptiles: Turtles follow it, lizards and snakes reverse it. Evolution 57: 1151 116. Atkinson D, 1994. Temperature and organism size: A biological law for ectotherms? Advances in Ecological Research 25: 1 58. Beaupre SJ, 1995. Effects of geographically variable thermal environment on bioenergetics of mottled rock rattlesnakes. Ecology 76: 1655 1665. Belk MC, Houston DD, 2002. Bergmann s rule in ectotherms: A test using freshwater fishes. American Naturalist 160: 80 808. Bergmann C, 1847. Ueber die verhältnisse der wärmeökonomie der thiere zu ihrer grösse. Göttinger Studien : 595 708. Brana F, Ji X, 2000. Influence of incubation temperature on morphology, locomotor performance, and early growth of hatchling wall lizards Podarcis muralis. Journal of Experimental Zoology 286: 422 4. Brandt R, Navas CA, 201. Body size variation across climatic gradients and sexual size dimorphism in Tropidurinae lizards. Journal of Zoology 290: 192 198. Brown RP, Griffin S, 2005. Lower selected body temperatures after food deprivation in the lizard Anolis carolinensis. Journal of Thermal 8

Biology 0: 79 8. Cruz FB, Fitzgerald LA, Espinoza RE, Schulte II JA, 2005. The importance of phylogenetic scale in tests of Bergmann's and Rapoport's rules: Lessons from a clade of South American lizards. Journal of Evolutionary Biology 18: 1559 1574. Cushman JH, Lawton JH, Manly BFJ, 199. Latitudinal patterns in European ant assemblages: Variation in species richness and body size. Oecologia 95: 0 7. Dawson WR, 1975. On the physiological significance of the preferred body temperatures of reptiles. Perspectives of Biophysical Ecology 12: 44 47. De Queiroz A, Ashton KG, 2004. The phylogeny of a species level tendency: Species heritability and possible deep origins of Bergmann's Rule in tetrapods. Evolution 58: 1674 1684. Du WG, Shine R, 2008. The influence of hydric environments during egg incubation on embryonic heart rates and offspring phenotypes in a scincid lizard Lampropholis guichenoti. Comparative Biochemistry and Physiology Part A, Molecular & Integrative Physiology 151: 102 107. Du WG, Xiang Ji, Zhang YP, Xu XF, Shine R, 2005. Identifying sources of variation in reproductive and lifehistory traits among five populations of a Chinese lizard (Takydromus septentrionalis, Lacertidae). Biological Journal of Linnean Society 85: 44 45. Dunham AE, Grant BW, Overall KL, 1989. Interfaces between biophysical and physiological ecology and the population ecology of terrestrial vertebrate ectotherms. Physiological Zoology 62: 5 55. Espinoza RE, Wiens JJ, Tracy CR, 2004. Recurrent evolution of herbivory in small, coldclimate lizards: Breaking the ecophysiological rules of reptilian herbivory. Proceedings of the National Academy of Sciences (USA) 101: 16819 16824. Feldman A, Meiri S, 2014. Australian snakes do not follow Bergmann s rule. Evolutionary Biology 41: 27 5. Flatt T, Shine R, BorgesLandaez PA, Downes SJ, 2001. Phenotypic variation in an oviparous montane lizard Bassiana duperreyi: The effects of thermal and hydric incubation environments. Biological Journal of the Linnean Society 74: 9 50. Gaston K, Blackburn T, 2008. Pattern and Process in Macroecology. Hoboken: WileyBlackwell. Grant BW, 1990. Tradeoffs in activity time and physiological performance for thermoregulating desert lizards Sceloporus merriami. Ecology 71: 22 2. Grant BW, Dunham AE, 1990. Elevational covariation in environmental constraints and life histories of the desert lizard Sceloporus merriami. Ecology 71: 1765 1776. Hailey A, Davies PMC, 1988. Activity and thermoregulation of the snake Natrix maura. 2. A synoptic model of thermal biology and the physiological ecology of performance. Journal of Zoology 214: 25 42. Hammond KA, Szewczak J, Król E, 2001. Effects of altitude and temperature on organ phenotypic plasticity along an altitudinal gradient. Journal of Experimental Biology 204: 1991 2000. 9

Heath JE, 1965. Temperature Regulation and Diurnal Activity in Horned Lizards. Berkeley: University of California Press. Hellmich WC, 1951. On ecotypic and autotypic characters, a contribution to the knowledge of the evolution of the genus Liolaemus (Iguanidae). Evolution 5: 59 69. Hu J, Xie F, Li C, Jiang J, 2011. Elevational patterns of species richness, range and body size for spiny frogs. PLoS ONE 6: e19817. Huang ZH, Liu NF, 2002. Evolution of the female reproductive strategy of Phrynocephalus vlangalii. Journal of Lanzhou University (Natural Sciences) 8:99 10. Ibargüengoytía NR, 2005. Field, selected body temperature and thermal tolerance of the syntopic lizards Phymaturus patagonicus and Liolaemus elongatus (Iguania: Liolaemidae). Journal of Arid Environments 62: 45 448. Janzen FJ, Ast JC, Paukstis G.L, 1995. Influence of the hydric environment and clutch on eggs and embryos of two sympatric map turtles. Functional Ecology 9: 91 922. Jin YT, Brown RP, 201. Species history and divergence times of viviparous and oviparous Chinese toadheaded sand lizards (Phrynocephalus) on the QinghaiTibetan Plateau. Molecular Phylogenetics and Evolution 68: 259 268. Jin Y, Liu N, Li J, 2007. Elevational variation in body size of Phrynocephalus vlangalii in the north QinghaiXizang (Tibetan) Plateau. Belgian Journal of Zoology 17: 197 202. Jin YT, Li JQ, Liu NF, 201. Elevationrelated variation in life history traits among Phrynocephalus lineages on the Tibetan Plateau: do they follow typical squamate ecogeographic patterns? Journal of Zoology 290: 29 01. Körner C, 2007. The use of altitude in ecological research. Trends in Ecology & Evolution 22: 569 574. Lindstedt SL, Boyce MS, 1985. Seasonality, fasting endurance, and body size in mammals. American Naturalist 125: 87 878. MartinVallejo J, GarcíaFernández J, PérezMellado V, VicenteVillardón J, 1995. Habitat selection and thermal ecology of the sympatric lizards Podarcis muralis and Podarcis hispanica in a mountain region of central Spain. Herpetological Journal 5: 181 188. Mayr E, 1956. Geographical character gradients and climatic adaptation. Evolution 10: 105 108. Mousseau TA, 1997. Ectotherms follow the converse to Bergmann's Rule. Evolution 51: 60 62. McCoy MW, Bolker BM, Osenberg CW, Miner BG, Vonesh JR, 2006. Size correction: Comparing morphological traits among populations and environments. Oecologia 148: 547 554. Muth A, 1977. Thermoregulatory postures and orientation to the sun: A mechanistic evaluation for the zebratailed lizard Callisaurus draconoides. Copeia 1977: 710 720. Nevo E, 1981. Genetic variation and climatic selection in the lizard Agama stellio in Israel and Sinai. Theoretical and Applied Genetics 60: 69 80. Partridge L, Coyne JA, 1997. Bergmann's rule in ectotherms: Is it adaptive? Evolution 51: 62 65. Palkovacs EP, 200. Explaining adaptive shifts in body size on islands: A life history approach. Oikos 10: 7 44. 10

Pianka ER, Vitt LJ, 200. Lizards: Windows to the Evolution of Diversity. Berkeley: University of California Press. PincheiraDonoso D, Meiri S, 201. An intercontinental analysis of climatedriven body size clines in reptiles: No support for patterns, no signals of process. Evolutionary Biology, DOI 10.1007/s11692 01 922 9. PincheiraDonoso D, Hodgson DJ, Tregenza T, 2008. The evolution of body size under environmental gradients in ectotherms: Why should Bergmann's rule apply to lizards? BMC Evolutionary Biology 8: 68. PincheiraDonoso D, Tregenza T, Hodgson DJ, 2007. Body size evolution in South American Liolaemus lizards of the boulengeri clade: A contrasting reassessment. Journal of Evolutionary Biology 20: 2067 2071. Roitberg ES, Kuranova VN, Bulakhova NA, Orlova VF, Eplanova GV et al., 201. Variation of reproductive traits and female body size in the most widelyraning terrestrial reptile: Testing the effects of reproductive mode, lineage, and climate. Evolutionary Biology 40: 420 48. Sears MW, Angilletta MJ Jr, 2004. Body size clines in Sceloporus lizards: Proximate mechanisms and demographic constraints. Integrative and Comparative Biology 44: 4 442. Shine R, 1980. Costs of reproduction in reptiles. Oecologia 46: 92 100. Stillwell RC, Fox CW, 2009. Geographic variation in body size, sexual size dimorphism and fitness components of a seed beetle: Local adaptation versus phenotypic plasticity. Oikos 118: 70 712. Tapponnier P, Xu Z, Roger F, Meyer B, Arnaud N et al., 2001. Oblique stepwise rise and growth of the Tibet Plateau. Science 294: 1671 1677. Van den Thillart G, Van Lieshout G, Storey K, Cortesi P, De Zwaan A, 1992. Influence of longterm hypoxia on the energy metabolism of the haemoglobincontaining bivalve Scapharca inaequivalvis: Critical O 2 levels for metabolic depression. Journal of Comparative Physiology B162: 297 04. Van Voorhies WA, 1996. Bergmann size clines: A simple explanation for their occurrence in ectotherms. Evolution 50: 1259 1264. Van Voorhies WA, 1997. On the adaptive nature of Bergmann size clines: A reply to Mousseau, Partridge and Coyne. Evolution 51: 65 640. Wang YZ, Zeng XM, Fang ZL, Wu GF, Papenfuss TJ et al., 1999. Study on the relationships of classification, phylogenetics and distribution of the genus Phrynocephalus spp. (Sauria: Agamidae) with the paleogeographical changes during Cenozoic era in Tibet Plateau. Zoological Research 20: 178 185 (in Chinese with English abstract). Volynchik S, 2014. Climaterelated variation in body dimensions within four lacertid species. International Journal of Zoology http://dx.doi.org/10.1155/2014/79587. Wu P, Wang Y, Wang S, Zeng T, Cai H et al., 2002. The age structure and sex ratio of Phrynocephalus vlangalii (Sauria: Agamidae). Journal of Sichuan University (Natural Science) 9: 118 119 (in Chinese with English abstract). Wu P, Wang Y, Zhu B, Zeng Z, 2004. Phrynocephalus vlangalii at Zoige, Sichuan: Burrow density and depth and their implications. 11

Zoological Research 25: 11 15 (in Chinese with English abstract). Zhang XD, Ji X, Luo LG, Gao JF, Zhang L, 2005. Sexual dimorphism and female reproduction in the Qinghai toadheaded lizard Phrynocephalus vlangalii. Acta Zoologica Sinica 51: 1006 1012. Zhao EM, Zhao KT, Zhou KY, 1999. Fauna Sinica, Reptilia, Volume 2, Squamata, Lacertilia. Beijing: Science Press (in Chinese). 12

Fig. 1 Linear and nonlinear regression analyses of mean snoutvent lengths (SVL; mm) of Phrynocephalus theobaldi and the first mean principle component scores (vertical axis) on elevation (horizontal axis, m). A total of 11 curve estimation models were tested using SPSS v19.0. Nonlinear analyses were conducted and produced the smallest AIC values when using quadratic and cubic models. 1

Table 1 Morphological measurements of P. theobaldi collected in 2011 from different elevations within Xizang portion of the Tibetan Plateau, southwestern China. All measurements (mm) are reported as mean ±1 SD. SVL=snoutvent length, TL=tail length, HL=head length, HW=head width, AL=arm length, LL=leg length, DBA=distance between axillae, DBI=distance between iliac crests and abdominal blackspeckled area (ABA). Elevation (m) n SVL TL HL HW AL LL DBA DBI ABA 600 6 48.27 ±.64 47.41 ±.15 11.88 ± 0.74 9.87 ± 0.6 12.76 ± 1.29 17.68 ± 1.11 9.6 ± 1.00 6.17 ± 0.74 0 650 17 47.76 ±.50 46.86 ±.06 12.12 ± 0.57 11.01 ± 0.6 1.12 ± 0.92 17.98 ± 2.6 10.40 ± 0.78 6.00 ± 0.67 0 800 48.22 ± 1.70 49.54 ± 2.7 11.92 ± 0.52 9.65 ± 0.62 12.45 ± 0.46 16.5 ± 1.05 9.86 ± 1.67 6.01 ± 0.45 0 900 16 48.78 ± 2.91 49.8 ±.87 12.65 ± 1.99 10.76 ± 0.68 12.56 ± 1.61 18.06 ± 1.68 10.88 ± 1.05 5.67 ± 0.66 0 4000 47.91 ± 1.98 5.2 ± 2.48 11.60 ± 1.49 10.24 ± 0.48 14.0 ± 0.25 19.55 ± 1.0 9.52 ± 0.0 5.80 ± 0.76 0 4050 8 46.77 ±.64 44.54 ± 2.01 11.58 ± 0.56 10.05 ± 0.55 12.41 ± 1.00 17.06 ± 0.71 10.46 ± 0.71 6.16 ± 0.51 0 4250 6 46.75 ± 2.94 52.9 ± 4.07 11.40 ± 0.24 9.7 ± 0.62 1.82 ± 0.66 20.1 ± 1.47 9.61 ± 0.77 5.56 ± 0.65 92.27±18.18 400 69 48.45 ±.47 54. ± 5.17 11.57 ± 0.98 10.09 ± 1.05 14.20 ± 1.22 20.17 ± 1.50 10.4 ± 1. 6.06 ± 1.07 89.5±28.09 450 50 49.62 ± 4.50 49.94 ± 4.2 11.72 ± 0.86 10.56 ± 0.91 1.8 ± 2.10 18.74 ± 1.79 10.81 ± 0.99 5.82 ± 0.52 66.61±2.27 4400 24 47.27 ±.2 52.49 ± 6.6 11.64 ± 0.64 10.27 ± 0.64 1.69 ± 1.17 19.48 ± 1.70 10. ± 1.02 5.91 ± 1.02 100.26±9.40 4450 46.0 ± 2.76 51.66 ± 10.65 11.98 ± 1.77 10.70 ± 1.45 1.78 ± 1.62 21.6 ± 1.86 10.7 ± 1.58 5.80 ± 0.89 115.62±12.10 4500 8 47.0 ±.64 52.00 ± 4.46 11.77 ± 0.97 10.07 ± 0.7 14.85 ± 1.54 19.85 ± 2.98 10.9 ± 0.97 6.04 ± 1.08 116.12±42.95 4550 4 50.06 ±.78 54.19 ± 5.87 12.4 ± 1.62 10.8 ± 1.09 14.14 ± 1.1 20.86 ± 1.74 11.15 ± 1.15 5.71 ± 0.51 108.89±5.97 4600 5 50.55 ± 4.00 5.8 ± 4.45 12.05 ± 0.72 10.81 ± 0.7 14.41 ± 1.02 21.18 ± 1.66 11.26 ± 1.5 5.99 ± 0.67 90.64±2.07 4650 0 50.70 ± 4.7 5.58 ± 5.27 12.22 ± 0.92 10.98 ± 0.90 14.7 ± 1.02 21.50 ± 1.64 11.1 ± 1.15 5.84 ± 0.78 89.95±4.20 4700 8 50.5 ± 4.71 50.01 ± 6.05 11.71 ± 0.54 10.4 ± 0.78 1.5 ± 1.21 19.52 ± 1.76 11.29 ± 1.49 6.4 ± 1.04 12.67±5.68 4750 7 47.59 ± 2.65 52.40 ± 5.25 11.79 ± 0.70 10.11 ± 0.8 14.55 ± 0.82 21.49 ±.71 10.82 ± 0.90 5.59 ± 0.5 111.81±40.26 4850 52.04 ± 4.00 51.1 ±.67 12.71 ± 0.6 11.4 ± 0.97 15.64 ± 0.45 22.8 ± 1. 11.81 ± 0.47 7.26 ± 1.40 109.26±8.209 5050 2 51.0 ± 0.1 50.04 ± 0.08 11.94 ± 0.01 10.46 ± 0.08 14.74 ± 0.06 20.21 ± 0.29 1.7 ± 0.0 7.44 ± 0.0 11.64±0.00 All sites 50 49.1 ±.96 52.16 ± 5.41 11.92 ± 1.04 10.52 ± 0.9 1.89 ± 1.44 19.9 ± 2.15 10.77 ± 1.2 5.94 ± 0.81 94.12±4.79 14

Table 2 The first principle component (PC) score of male and female P. theobaldi at different elevations (values are reported as mean ±1 SD). Elevation (m) Male Female PC1 n PC1 n 600 0.789 ± 0.0 1 0.462±0.624 5 650 0.698±0.10 6 0.026±1.048 11 800 0.684±0.28 900 0.48±1.12 10 0.162±0.864 6 4000 0.799±0.9 2 0.17±0.000 1 4050 0.448±0.020 2 0.696±0.760 6 4250 0.950±0.48 0.64±0.72 400 0.547±1.142 6 0.069±1.104 450 0.467±0.759 2 0.45±0.96 26 4400 0.540±0.895 10 0.8±0.646 14 4450 0.54±1.14 4500 0.08±1.162 4 0.102±0.89 4 4550 0.525±0.975 25 0.284±0.85 9 4600 0.124±0.716 29 0.467±0.945 24 4650 0.19±0.999 15 0.519±1.146 15 4700 0.74±0.000 2 0.154±0.92 6 4750 0.541±0.822 0.289±0.999 4 4850 0.469±0.000 1 1.451±0.98 2 5050 1.021±0.067 2 All sites 175 175 15

Table Loading values and the percentage of total variance explained for each of the first two principle components (PC) among eight morphological traits of males (n =175) and females (n =175) for P. theobaldi. Character Female Male PC1 PC2 PC1 PC2 SVL 0.878 0.102 0.88 0.15 TL 0.628 0.51 0.688 0.284 HL 0.804 0.25 0.61 0.557 HW 0.795 0.6 0.82 0.02 AL 0.671 0.476 0.55 0.66 LL 0.71 0.506 0.651 0.548 DBA 0.81 0.208 0.828 0.201 DBI 0.51 0.77 0.59 0.029 Total variance 54.01 14.67 50.67 15.65 16

Table 4 Linear and nonlinear quadratic and cubic regressions of female snoutvent length (SVL) and PC1 against elevation for P. theobaldi (for all comparisons, df = 2, 16). Dependentindependent Model r 2 F P Constant Coefficient b1 Coefficient b2 Coefficient b AIC Female SVLelevation Quadratic 0.410 5.21 0.019 16.021 0.0427 5.2E06 17.284 Cubic 0.409 5.190 0.0190. 104.045 0.020 0 4.0E10 17.18 Linear 0.172.4 008 41.471 0.0018 2.77 Female PC1elevation Quadratic 0.407 5.14 0.020 44.172 0.0210 2.5E06 18.566 Cubic 0.422 5.48 0.016 14.6196 0 2.0 E06.8E10 19.040 Linear 0.01 0.21 0.654 0.6757 0.0002 9.98 Male SVLelevation Quadratic 0.8 4.5 0.04 2.48 0.04 4.0E06 25.996 Cubic 0.8 4.5 0.04 11.85 0.0191 25.984 Linear 0.42 7.78 0.014 1.4716 0.007 27.102 Male PC1elevation Quadratic 0.174 1.48 0.262 1.6924 0.0071 9.0E07 21.540 Cubic 0.176 1.48 0.259.811 8.0E07 1.4E10 21.569 Linear 0.118 2.01 0.177 2.185 0.0005 20.42 17

Table 5 Pearson correlation analyses (with twotailed significance values) of SVL and PC1 of P. theobaldi with climatic factors. Correlation coefficient (Sig. (two tailed)) Temperature Barometric pressure Sunshine hours Female SVL(>4200m) 0.769(0.00) 0.769(0.00) 0.769(0.00) Female PC1(>4200m) 0.749(0.005) 0.749(0.005) 0.749(0.005) Female SVL(<4200m) 0.040(0.941) 0.040(0.941) 0.040(0.941) Female PC1(<4200m) 0.160(0.762) 0.160(0.762) 0.160(0.762) Male SVL 0.574(0.016) 0.574(0.016) 0.574(0.016) Male PC1 0.6(0.188) 0.6(0.188) 0.6(0.188) 18