Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Swimming with Predators and Pesticides: How Environmental Stressors Affect the Thermal Physiology of Tadpoles

  • Marco Katzenberger ,

    Affiliation Department of Evolutionary Ecology, Doñana Biological Station - Spanish Council for Scientific Research, Sevilla, Spain

  • John Hammond,

    Affiliation Department of Biology, University of New Mexico, Albuquerque, New Mexico, United States of America

  • Helder Duarte,

    Affiliation Department of Evolutionary Ecology, Doñana Biological Station - Spanish Council for Scientific Research, Sevilla, Spain

  • Miguel Tejedo,

    Affiliation Department of Evolutionary Ecology, Doñana Biological Station - Spanish Council for Scientific Research, Sevilla, Spain

  • Cecilia Calabuig,

    Affiliation Department of Animal Sciences, Federal Rural University of the Semiarid Region, Mossoró, Rio Grande do Norte, Brazil

  • Rick A. Relyea

    Affiliation Department of Biological Sciences, University of Pittsburgh, Pittsburgh, Pennsylvania, United States of America

Swimming with Predators and Pesticides: How Environmental Stressors Affect the Thermal Physiology of Tadpoles

  • Marco Katzenberger, 
  • John Hammond, 
  • Helder Duarte, 
  • Miguel Tejedo, 
  • Cecilia Calabuig, 
  • Rick A. Relyea


To forecast biological responses to changing environments, we need to understand how a species's physiology varies through space and time and assess how changes in physiological function due to environmental changes may interact with phenotypic changes caused by other types of environmental variation. Amphibian larvae are well known for expressing environmentally induced phenotypes, but relatively little is known about how these responses might interact with changing temperatures and their thermal physiology. To address this question, we studied the thermal physiology of grey treefrog tadpoles (Hyla versicolor) by determining whether exposures to predator cues and an herbicide (Roundup) can alter their critical maximum temperature (CTmax) and their swimming speed across a range of temperatures, which provides estimates of optimal temperature (Topt) for swimming speed and the shape of the thermal performance curve (TPC). We discovered that predator cues induced a 0.4°C higher CTmax value, whereas the herbicide had no effect. Tadpoles exposed to predator cues or the herbicide swam faster than control tadpoles and the increase in burst speed was higher near Topt. In regard to the shape of the TPC, exposure to predator cues increased Topt by 1.5°C, while exposure to the herbicide marginally lowered Topt by 0.4°C. Combining predator cues and the herbicide produced an intermediate Topt that was 0.5°C higher than the control. To our knowledge this is the first study to demonstrate a predator altering the thermal physiology of amphibian larvae (prey) by increasing CTmax, increasing the optimum temperature, and producing changes in the thermal performance curves. Furthermore, these plastic responses of CTmax and TPC to different inducing environments should be considered when forecasting biological responses to global warming.


Biological mechanisms underlying a response to environmental changes can be quite complex. To forecast these biological responses, we need to understand how a species' physiology varies through space and time [1], [2] and assess how changes in physiological function induced by environmental changes (e.g., increasing environmental temperatures) may interact with phenotypic changes induced by other types of environmental variation [3], [4], [5], [6].

Species can possess the ability to respond to new or altered environments with flexible phenotypes that are environmentally induced and can potentially contribute to adaptive evolution [7]. Stressful environments can induce non-adaptive plasticity, increasing the variance around the mean phenotypic response or distancing it from the favored optimum. Nevertheless, if plasticity is adaptive and promotes establishment and persistence in a new environment, by placing populations close enough to a new phenotypic optimum for directional selection to act, it can predictably enhance fitness and is most likely to facilitate adaptive evolution on ecological timescales [7].

The presence of predators in the environment can induce behavioral and morphological changes in prey that result in the prey being less susceptible to the predator (e.g., [8], [9], [10], [11]). Furthermore, pesticides can also induce behavioral and morphological changes in organisms. Sublethal exposure to pesticides early in life can make the individuals more tolerant of the pesticide later in life [12], [13] and they can induce phenotypic changes that resemble predator-induced phenotypes [14], [15], [16], [17]. In other cases, pesticides impede the induction of predator-induced morphology [18], [19], [20], [21].

In the current scenario of climate change, there has been a renewed interest in the thermal physiology of organisms and the estimation of thermal tolerance and sensitivity, using physiological traits such as the critical thermal maximum (CTmax; e.g., the temperature at which animals become immobile [22], [23]), the optimum temperature (Topt) for performing some function, or the shape of the thermal performance curve (TPC), which describes how an animal's performance changes across a range of temperatures. Although some pesticides are known to affect CTmax and burst speed, usually in a negative way (e.g., [24]), there is limited information on how pesticides affect optimum temperature and performance over a range of temperatures (i.e. how pesticides affect TPCs), especially for amphibians. Likewise, much is known about predator-induced changes in organisms, including some interactions with pesticides [17]. Predators also influence thermoregulation and thermal preferences of prey, resulting in behavioral changes and coevolution of thermal optima between species [25]. Other than these behavioral responses that indirectly affect physiology, little is known about whether predator cues can directly affect the thermal physiology of prey.

We addressed these issues by studying the thermal physiology of grey treefrog tadpoles (Hyla versicolor LeConte 1825) that were exposed to predator cues and pesticides. Tadpoles are excellent model organisms for this study because they are practically isothermal with their aquatic environment [23] and their thermal physiology traits (CTmax and Topt) are not influenced by confounding processes such as dehydration. Tadpoles are also well known for expressing predator-induced changes in behavior and morphology (e.g., [9], [26], [27]. Furthermore, at least two species of tadpoles can alter their morphology when exposed to the herbicide Roundup and exhibit morphological changes that closely resemble predator-induced changes in tadpoles [17].

Given that pollutants and predators can both affect many aspects of tadpole biology, including development and metamorphosis (e.g., [28], [29]), and the interaction of pollutants with other stressors are often negative to the organism (e.g., glyphosate, [30]), we expect the impact of these stressors on the thermal physiology of tadpoles to be mainly negative. Therefore, we hypothesized that tadpoles exposed a sublethal concentration of an herbicide will have reduced tolerance to higher temperatures (CTmax) and exhibit a lower optimal temperature (Topt) compared to tadpoles not exposed to the herbicide. Furthermore, because predator cues and the herbicide can induce deeper tails in tadpoles, we hypothesized that tadpoles exposed to either stressor will suffer a vertical shift upward in their TPC across a range of temperatures [31], and have increased swimming performance (e.g., [32]). However, it is also possible that the herbicide will have a negative effect on swimming performance (e.g., [33]) if induced morphological changes are countered by other phenotypic changes that impair swimming ability.


Inducing the tadpoles

The induction experiment was conducted at the University of Pittsburgh's Pymatuning Laboratory of Ecology in northwest Pennsylvania, USA. The experiment used a completely randomized, 2×2 factorial design comprised of the presence or absence of predator cues crossed with the presence or absence of an herbicide (nominal concentrations of 0 or 2 mg active ingredient per liter (a.e./L). Based on past studies, this herbicide concentration should remain sublethal to gray treefrog tadpoles while inducing morphology changes (e.g., [34], [35]).

The four treatment combinations were replicated four times for a total of 16 mesocosms, which consisted of 120-L wading pools, set outdoors (air temperature ranged from 9°C to 28°C), that we filled with 100 L of well water on 11 June 2011. We then added 100 g of dry leaves (Quercus spp.) and 5 g of rabbit chow to serve as habitat structure and an initial nutrient source, respectively. We also added an aliquot of zooplankton and phytoplankton that was a mixture from 5 local ponds. Each mesocosm was equipped with a predator cage constructed of 10×10 cm well pipe covered with window screen at each end. These cages allow the chemical cues emitted during predation to diffuse through the water while preventing the predators from killing the target tadpoles [36], [37], [38]. Mesocosms were covered with a 60% shade cloth, for the duration of the outdoor experiment.

To obtain tadpoles for the experiment, we collected >20 amplecting pairs of grey treefrogs from a nearby wetland (41° 34′ 9.55" N, 80° 27′ 22.29" W) on 18, 21 and 22 May 2011, and allowed them to lay eggs in tubs containing aged well water. Once the eggs hatched, the tadpoles were held in outdoor pools and fed rabbit pellets ad libitum until used in the experiment.

On 15 June 2011, which we defined as day 0 of the experiment, we added 40 tadpoles to each mesocosm from a mixture of the clutches with an initial mass (±SE) of 37.5±2.1 mg per tadpole (subsample, N = 20). On day 1, we applied the herbicide treatment. To achieve nominal concentrations of 2 mg a.e./L, we prepared 8 equal mixtures containing 372 µL of stock solution (Roundup Power Max; concentration  = 540 g a.e./L) and 250 ml of water. For the eight mesocosms assigned the herbicide treatment, we drizzled one mixture into each mesocosm. For the eight mesocosms assigned the no-herbicide treatment, we drizzled 250 mL of water into each mesocosm. Approximately 1 hr after dosing, we collected water samples from each tank to confirm the concentration of the herbicide. An independent analysis found that the concentrations in the water were 0 and 1.55 mg a.e./L (Mississippi State Chemical Laboratory, Mississippi State, MS). Observing lower actual concentrations is a common phenomenon in mesocosm experiments (reviewed in Brock et al. 2000), likely as the result of binding to surfaces in the mesocosm and degradation of the samples before the testing is conducted. Jones et al. [39] measured little herbicide breakdown for a similar time period, so we assumed there was little change in herbicide concentration during the induction experiment.

After sampling the water, we manipulated the predator environment. For mesocosms assigned the no-predator treatment, the cages remained empty. For mesocosms assigned the predator-cue treatment, we placed a single dragonfly nymph (Anax junius) inside the predator cage. Each dragonfly was fed ∼300 mg of grey treefrog tadpole biomass every 2 d (see [38]). Prior to each feeding, we observed no tadpoles left in the predator cage, which indicates that the dragonfly nymphs consumed the tadpoles in the cages. The feeding continued until day 10 to allow tadpole growth and induction by the herbicide and predator cues.

Determining the critical thermal maximum of the tadpoles

On day 10, we brought sets of tadpoles into the laboratory to allow them to acclimate at a temperature of 20°C (approximately the average temperature experienced in the mesocosms), with a 12L:12D photoperiod, for 4 to 5 d before testing them for CTmax and Topt [22], [40]. During acclimation, tadpoles were fed rabbit pellets ad libitum and we maintained the predator and herbicide environments to help prevent the loss of any phenotype induction [41]. All tested larvae were below Gosner stage 38 [42]. This is important because tadpoles close to metamorphic climax exhibit a significant decline in thermal tolerance [43].

We obtained upper critical thermal tolerances (CTmax) by using a slightly modified version of Hutchison's dynamic method [23]. We exposed tadpoles to a constant heating rate of 0.05°C min−1 (3°C h−1), which simulates a natural rate of temperature increase in ponds (H. Duarte, M. Tejedo, J. Hammond, M. Katzenberger, R.A. Relyea, unpublished data from dataloggers; see also [44]) until we observed complete immobility, which signaled the endpoint of the experiment. After reaching CTmax, we transferred tadpoles to cooler water (∼20°C) to allow recovery. After complete recovery, the tadpoles were weighed and we found that the mass of the tadpoles had increased by 13- to 15-fold since day 0. We tested 3 to 4 tadpoles from each mesocosm, for a total of 56 tadpoles from the 16 mesocosms, as seen in Table 1.

Table 1. Critical thermal maximum (CTmax), sample size (N) and body mass (Mass) of Hyla versicolor tadpoles, in four treatments.

We performed an analysis of variance (ANOVA) that used CTmax as the dependent variable, predator cues and herbicide as categorical factors (including the interaction of these factors), and mesocosm nested within the interaction of predator cues and herbicide (i.e. mesocosm nested within treatment). Given that tadpole mass was not correlated with CTmax (see results), we did not include it as a covariate. No data transformations were required for this analysis.

Determining the thermal performance curves for tadpole burst speed

Locomotor performance, measured as a TPC, is considered to be a proxy of maximum physiological performance and has been used to estimate optimum temperatures in amphibians [45], [46]. We obtained TPCs by measuring each tadpole's maximal burst swimming speed (i.e. burst speed) across a range of temperatures. To determine burst speed, tadpoles were placed individually in a portable thermal bath (patent license ES 2372085), which consists of an opened cross section methacrylate tube (1 m long ×6 cm wide ×3 cm deep) filled with water of a given temperature. We then gently prodded the tadpole with a thin stick to stimulate swimming. Each trial was recorded using a digital camera (30 frames/s) positioned above the tube (JVC Everio GZ-MG505). TPCs were defined using a set of six temperatures (20°, 24°, 28°, 32°, 35° and 38°C). This set includes temperatures tadpoles experienced in the mesocosms (20°-32°C) and two more (35° and 38°C) which they might be exposed to in a scenario of increasing environmental temperatures (but lower than their critical thermal maximum). Temperatures were tested in a random order and, for each temperature, tadpoles from the four treatments were tested in the same session; therefore, all treatments had the same temperature order. Prior to swimming, tadpoles were held individually in 250-ml containers at the test temperature for approximately 1 hr. A different set of tadpoles (total N = 570) was used for each temperature (Table 2) and each wading pool was represented equally in each set.

Table 2. Experimental temperatures, sample size (N), total tadpole length in mm (TTL±SE), and maximum swimming speeds in cm/s (mean ±SE) for gray treefrog tadpoles that were exposed to predator cues and the herbicide Roundup.

After the tadpole started to move, we used the software Measurement in Motion [47] to estimate burst speed over three frames (0.1 s) by measuring the distance the center of mass moved between frames [48], [49]. After conducting at least three bouts, we used the fastest speed measured for a given tadpole as our measure of that individual's burst speed. Since maximal swimming speed may scale with body size [45] and body size may confound the effect of speed on escape success [50], we used size-corrected burst speed (using tadpole total length) when constructing TPCs.

To describe the TPCs for burst speed, we used the Template Mode of Variation method (TMV, [51]) which employs a polynomial function to decompose variation among TPCs into three predetermined modes of variation with biological connotation: vertical shift (faster-slower), horizontal shift (hotter-colder), and specialist-generalist trade-offs ([31]; see [51] and supporting information for details on calculations). Since we tested tadpole performance at six temperatures, we assumed that the common template curve was a fourth-degree polynomial, as in previous studies (e.g., [46]). Making this assumption avoids inadequately describing TPCs, which can happen when using a lower-order polynomial [51], [52].

In addition to using the TMV method, we also calculated maximum performance (zmax) to evaluate changes in maximum swimming speed at the optimum temperature and a more traditional measurement of performance breadth to confirm specialist-generalist trade-offs (using B95, which is the range of temperatures at which performance values exceed 95% of the maximum;[53]). We used B95 instead of the traditional B80 because the lower limit of B80 would fall below 20°C, which is outside the tested range of temperatures. All computations regarding the TPCs, except for B95, were made using the Matlab code by R. Izem (available online in the appendix of [51]). We also confirmed the fit of each treatment's curve and calculated standard error (SE) of each curve's parameters using nlinfit and nlparci functions, respectively, in Matlab [54].

We conducted an ANOVA that used burst speed as the dependent variable, temperature, predator cues and the herbicide (including the interaction of these factors) as categorical factors and, mesocosm nested within the interaction of predator cues and herbicide (i.e. mesocosm nested within treatment). ANOVA analysis was followed by a Tukey post-hoc test.

Assessing the morphology of the tadpoles

After the swimming trials, we determined the mass and developmental stage of each tested tadpole. We then took lateral photos of each tadpole and digitized the images for morphometric measurements. We captured the shape of tadpoles by digitizing 10 landmarks and 15 semi-landmarks (see supporting information; see also [49], [55]) on each tadpole using tpsDig2 software [56]. We then extracted partial warps and the uniform component with tpsRelw software [57], which we used as our shape variables in a subsequent analysis. We visualized variation in landmark positions using the thin-plate spline approach (transformation grids, [58] in MorphJ [59]. As an alternative approach to quantify tadpole morphology, we also took the following linear measurements of each tadpole: total tadpole length (TTL, distance between snout and tip of tail fin), body length (BL, distance between snout and point where bottom edge of tail muscle meets body), body depth (BD, deepest point of the body), tail length (TL, distance between point where bottom edge of tail muscle meets body and tip of tail fin), muscle depth (MD, deepest point of the muscle) and tail depth (TD, maximum depth of the tail fin).

We conducted canonical correlation analysis as a dimension-reducing procedure to obtain two morphological indices (i.e. a linear combination of shape variables); one was for the linear measurements (MIlin) and the other was for the partial warps and uniform component (MIgeo). We then examined these two indices for correlations with burst speed (across all treatments; see [55]). To determine if predator cues, herbicide, and their interaction influenced tadpole size (i.e. centroid) or shape (MIlin or MIgeo), we performed three ANOVAs followed by Tukey HSD post-hoc tests; mesocosms were nested within the interaction of predator cues and herbicide (i.e. mesocosm nested within treatment). Shape variables (MIlin and MIgeo) and tadpole size (centroid) were then used as continuous predictors, along with temperature, predator cues and herbicide as a categorical predictors, in two ANCOVA analysis (testing either MIlin or MIgeo separately), to evaluate their effects on burst speed. We performed all analyses using Matlab [54], except when mentioned otherwise, and used a significance level of α = 0.05.

All experiments were approved by the University of Pittsburgh's Institutional Animal Care and Use Committee (Protocol #12050451).


Critical thermal maxima of the tadpoles

In our analysis of CTmax, there were no differences among mesocosms within a given treatment. We found an effect of predator cues but no effect of the herbicide or the interaction of both (Table 3). Averaged across herbicide treatments, tadpoles exposed to predators had a CTmax that was 0.4°C higher than tadpoles not exposed to predators (Table 1). CTmax was not correlated with tadpole mass (Pearson's R = −0.17, p = 0.22).

Table 3. ANOVA using CTmax as dependent variable, predator cues and Roundup as categorical factors (including the interaction of these factors) and, mesocosm nested within the interaction of predator cues and Roundup, for Hyla versicolor.

Thermal performance curves for tadpole burst speed

When we test tadpole swimming ability across different water temperatures, we found that swimming burst speed varied with temperature (Table 2). When we used the TMV method on size-corrected performance data, we obtained both a common template curve, which provided a good approximation of the common shape of each treatment's curve (Fig. 1), and a three-parameter shape-invariant model (with the use of a fourth-degree polynomial), which explained over 99% of the variation for swimming speed. Decomposition of the total variation into the three pre-determined directions of variation reveals that TPCs for swimming speed vary mostly in the specialist-generalist (53.27%) direction and the vertical (45.98%) direction, but very little in the horizontal (0.59%) direction. This indicates that tadpoles in the control treatment had a wider swimming TPC than tadpoles exposed to predator cues or the herbicide, even when comparing more traditional measures of curve width (B95; Table 4, Fig. 2). Thus, most of the variation in the TPCs is due to specialist-generalist trade-offs and differences in overall performance (faster-slower), rather than changes in Topt (hotter-colder). Indeed, tadpoles raised in the herbicide treatment exhibited only a small decrease in Topt (−0.4°C) while tadpoles raised with predator cues exhibited an increase in Topt (1.5°C). Tadpoles raised with both predators and herbicide exhibited a Topt that was intermediate in magnitude between the latter two treatments but still higher (0.5°C) than tadpoles raised in the control treatment. The only significant difference in Topt was between tadpoles exposed only to herbicide and those exposed only to predator cues (1.8°C; 2-tailed t-test, p<0.05). Maximal performance (zmax) was marginally correlated with performance breadth (Pearson's R = −0.95, p = 0.051).

Figure 1. Rescaled thermal performance curves for swimming speed in each treatment with fitted common template shape.

Common template shape z(T) is represented by a dashed line nad the treatments by solid lines. Each thermal performance curve of a treatment (i) and temperature were standardized with respect to the estimates of height (h), location (m; Topt), and width (w) parameters from the fit to model. Rescaled optimum temperature Topt = 0. (see [46], [51]). Swimming z(T)  = 1.6458–0.004T2–0.00023982T3+0.000003493T4.

Figure 2. Overall shape of the thermal performance curves for each of the four induction treatments.

Each treatment is represented by a thermal performance curve for tadpole swimming speed: control - solid line, predator - dashed line, Roundup - dotted line and predator+Roundup - dash-dot line.

Table 4. Parameters of thermal performance curves for maximum swimming speed in four treatments, for Hyla versicolor, estimated with TMV method (Izem and Kingsolver, 2005) and nlinfit/nlparci functions in Matlab (Mathworks, 2009).

Temperature and predator cues both influenced burst speed (Table 5). There was also a significant interaction between predator cues and herbicide. Tadpoles in the control treatment had slower burst speeds accross all temperatures than tadpoles in the other three treatments (all p<0.05). Tadpoles raised in the predator treatment were also faster than those from herbicide treatment (p<0.05). Furthermore, tadpoles in all treatments containing predator cues or herbicide had higher maximum performance (zmax) than tadpoles in the control treatment, so that their burst speed at the optimum temperature was higher than the burst speed of tadpoles raised without any cues. These differences in the parameters of the TPCs can be seen as changes in the overall shape of the curves (Fig. 2). Our analysis of burst speed also revealed a significant effect of mesocosms (nested within treatment), however the magnitude of this effect was much smaller than in other effects, such as the interaction of predator cues and herbicide (Table 5). Nevertheless, we checked for burst speed differences among tanks of the same treatment and temperature and we found no significant effect of mesocosm on burst speed, in any of the treatment-temperature combinations (all p>0.05).

Table 5. ANOVA using burst speed as dependent variable, and temperature, mesocosm, predator cues and Roundup as categorical predictors, with mesocosm nested within the interaction of predator cues and Roundup, for Hyla versicolor.

Induced morphology of the tadpoles

We observed size and shape changes in tadpoles exposed to the herbicide and predator cue treatments (Fig. 3). Predator cues and herbicide had no main effects on tadpole centroid size (Table 6a) but they did have a significant interaction; tadpoles exposed to predator cues + herbicide were smaller than those exposed only to the herbicide or only to the predator cues (both p<0.05). Similarly, tadpoles in the control treatment were smaller than those exposed only to the herbicide or only to the predator cues (both p<0.05). For geometric morphometric measurements, both predator cues and herbicide influenced tadpole shape (Table 6b) and there was a significant interaction between the two factors. Tadpoles raised in the control treatment differed from those raised in the other three treatments (all p<0.05), however these did not differ amongst themselves. For linear measurements, only predator cues significantly influenced shape of tadpoles (Table 6c). Tadpoles raised in predator or predator + herbicide treatment differed from those raised in herbicide or control treatments (all p<0.05). Mesocosm effect on either centroid or shape (MIlin or MIgeo) was non-significant (Table 6). Overall, compared to tadpoles in the control, tadpoles in the other three treatments exhibited relatively shorter bodies. Furthermore, in the two treatments containing predator cues, tadpoles exhibited an increase in their relative tail length and tail depth (Fig. 3). Apart from temperature and predator cues, burst speed was also influenced by tadpole's size, either when using morphometric geometric data (Table 7a) or linear measurements (Table 7b). We also found a significant effect of shape on burst speed when using geometric morphometric data (Table 7a).

Figure 3. Transformation grids with landmarks and warped outline drawings for each treatment's tadpole shape.

a) Transformation grids with landmarks (black dots) and vectors showing direction of variation; b) comparison of warped outline drawings for each treatment shape (black) and control shape (grey). Transformation grids and warped outline drawings were magnified (x5) to better illustrate the differences. C – Control, R – Roundup, P – Predator and PR – Predator + Roundup.

Table 6. ANOVAs to determine if predator cues and Roundup (including their interaction) influenced size (a; centroid), or shape (b and c) of tadpoles (MIgeo, for geometric morphometric measurements, or MIlin, for linear measurements, respectively) with mesocosm nested within the interaction of predator cues and Roundup (i.e. mesocosm nested within treatment).

Table 7. ANCOVA analysis using burst speed as dependent variable, shape variables MIgeo (a) or MIlin (b) and tadpole size (centroid) as continuous predictors, alongside temperature, predator cues and Roundup as categorical predictors.


We discovered that predator cues and the herbicide Roundup can affect the thermal physiology of Hyla versicolor tadpoles. Predator cues induced tadpoles to have CTmax values that were 0.4°C higher whereas the herbicide had no effect. Predator cues and Roundup also influenced the shape of the thermal performance curves, resulting in changes in optimum temperature, performance breadth and maximal performance (Fig. 2). Furthermore, predator cues also induced morphological changes that increased the tadpoles' burst speed.

Roundup, a glyphosate based broad-spectrum systemic herbicide, did not have any effect on CTmax estimates of tadpoles. However there have been reports of other contaminants affecting the thermal physiology of vertebrates. Among insecticides, for example, endosulfan (an organochlorine insecticide that affects the central nervous system) and chlorpyrifos (an organophosphate insecticide that inhibits acetylcholinesterase) are known to decrease CTmax in fishes [60]. Other environmental contaminants, such as cadmium and copper, can adversely affect the ability of fish to withstand high temperature stress [61], [62]. Whether all of these observations in fishes can be extrapolated to other species of aquatic organisms, such as tadpoles, is yet to be determined. Based on these studies and our own results, it seems that the effects of pesticides on CTmax may depend on the type of pesticide, the concentration of the pesticide, and how it affects the organism (i.e. its mode of action). There is the possibility that using higher concentrations of the herbicide might induce a decrease in CTmax, but higher concentrations will cause tadpole death [63]. Furthermore, the herbicide also did not interfere with the increase in CTmax induced by predator cues; tadpoles exposed to predator cues + herbicide had similar CTmax values to those exposed only to predator cues.

Different methodological protocols and biological sources can affect estimates of upper thermal tolerances (see [44], [64]). For example, the ramping rate used [65], [66], [67], [68], the selection of end-point [23], variations in previous thermal acclimation [40], ontogenetic stage [43], time of day, and photoperiod [69] all may promote shifts in amphibian upper thermal tolerances. We discovered that predatory cues can also affect CTmax estimates of prey. An increase in thermal tolerance of predator-induced tadpoles would cause an increase in their warming tolerance, which is the difference between CTmax and maximum temperature of the environment to which an ectotherm is exposed [70], [71]. This means that tadpoles exposed to predator cues would be less susceptible to acute thermal stress than tadpoles that were not exposed to predator cues. In contrast, an exposure to the herbicide, at least at the concentration used in our study, would not affect the warming tolerance of tadpoles.

An exposure to predator cues and the herbicide had interactive effects on tadpole burst speed. The interaction occurred because the herbicide alone and predator cues alone each increased burst speed compared to the control, but the combination of the herbicide and predator cues induced an increase that was not larger than predator cues alone.Therefore, since the combination of the herbicide and predators cues was not additive, in the presence of predator cues, exposure to the herbicide caused no change in burst speed.

The presence of either predator cues or the herbicide narrowed the performance breadth of the TPC while increasing maximal performance. As performance breadth is negatively correlated with maximal performance, we would expect a generalist-specialist trade-off. Tadpoles from a treatment which induced a more specialist curve (as demonstrated by predator cues + herbicide) would perform better at the optimum temperature but gradually decrease in performance, as moving away from the optimum temperature, until reaching a point were tadpoles from a treatment which induced a more generalist curve (as demonstrated by control) would outperform them (see [31], [51]; Table 4). However, we do not see a decline in performance at the extremes of the thermal performance curve, at the tested temperatures, as a result of this trade-off. This observation is confirmed by thermal tolerance data where none of the tadpoles raised in any of the treatments with predator cues or the herbicide had lower CTmax than those from the control treatment. Instead, it appears the expected decline in sub-optimal performance resulting from a generalist-specialist trade-off is compensated by the increase in overall performance, so that tadpoles raised in the control treatment always perform, on average, worse than herbicide- or predator-induced tadpoles, at least at the tested temperatures. Therefore, when comparing thermal performance curves, the resulting increase in overall performance was asymmetric, being greater around the optimum temperature and lower at the extreme temperatures.

Surprisingly, predator cues and the herbicide also produced changes in the optimum temperature, but in opposite directions. Of course, the small decrease in optimum temperature caused by the herbicide (0.4°C) may have little or no biological relevance. In contrast, the increase in optimum temperature promoted by predator cues (approximately 1.5°C) may be important, especially when new assessments suggested that environmental impacts will require smaller degrees of global warming than previously thought [72]. Since predator cues increase optimum temperature, the difference between optimum temperature and the environmental temperature should also increase (i.e. thermal safety margins (TSM); see [70]), which would be beneficial to the tadpoles in the current scenario of increasing global temperatures.

Previous studies have demonstrated that changes in the shape or position of thermal performance curves can occur due to acclimation (e.g., [73], [74], [75]) or that thermal performance curves of different locomotor strategies for the same organism can have different shapes (e.g., [46], [76]). In the present study, we demonstrate that the presence of sublethal concentrations of an herbicide and cues from predators can also produce changes in the thermal performance curves and therefore affect how tadpoles respond to environmental temperature changes.

Although it has been documented that predators can affect the behavioral thermoregulation of their prey (e.g., [25]), to our knowledge this is the first study to demonstrate a predator altering the thermal physiology of their prey by increasing CTmax, increasing the optimum temperature, and producing changes in the thermal performance curves. It has also been demonstrated that Roundup's lethality increases with competition stress [35] and that predator cues can improve tadpole survival when tadpoles are exposed to the herbicide under stratified water conditions [17]. Therefore, one could make the argument that acclimation to predator cues might be beneficial under warmer temperatures. However, we should also keep in mind that predation simultaneously has a negative effect on tadpole populations and can select for particular phenotypes (see [9]). To display a predator-induced phenotype, tadpoles need to detect chemical cues that are released when other tadpoles (particularly conspecifics) are consumed. So, the possible positive effects of predator cues on the thermal physiology, in a global warming scenario, would only be beneficial for those phenotypes that survive predation.

Predator cues in our study induced morphology changes (relative smaller bodies, deeper tails and deeper tail muscle) that were similar to those observed in previous studies (e.g., [77]). These morphological changes likely explain why tadpoles exposed to predator cues swam faster than control tadpoles. Exposure to the herbicide (see figure 3) induced relative smaller bodies, and the observed changes partially resembled the predator-induced phenotype (see also [17]). The induction of relatively deeper tadpole tails by the herbicide was less evident in the current work than in the study of Relyea [17]. However, this may be due to a number of differences in the experimental protocol including the duration of exposure and a substantially different experimental venue.

Predator cues and the herbicide caused interactive effects on tadpole size. Tadpoles exposed to predator cues + herbicide were smaller than those exposed only to the herbicide or only to predator cues. Tadpoles raised in the control treatment also tended to be smaller than those exposed only to the herbicide or only to predator cues. This may explain why tadpoles from the herbicide treatment also swam faster than tadpoles from the control treatment. As a result, all three treatments had better overall swimming performance than in control, with increase in burst speed related to the magnitude of morphology change (more induction, higher performance) and size. Furthermore, predator-induced morphology changes can be reversed if cues are removed [41]. As a result, some of the changes in the thermal performance curve may also be reversible. If so, in the absence of cues, the predator- and herbicide-induced TPC shapes would revert back to the original curve (i.e. the control curve).

The mechanism underlying the ability of the herbicide to induce morphological changes in tadpoles is still unknown. It has been suggested that the herbicide may be interfering with the stress hormones that induce anti-predator defenses [78] or that herbicides and predator cues activate shared endocrinological pathways [17]. We have demonstrated that predator cues and the herbicide can affect the thermal physiology of tadpoles, although not all changes occur in the same direction. However, the mechanisms behind these thermal physiology changes are also unknown, with possible scenarios arising from our results: a) herbicide interferes only with the stress hormones that induce anti-predator defenses; b) they do not share the same physiological pathways, or at least not all of them; c) they both activate shared endocrinological pathways but predator cues also indirectly activate temperature-stress response mechanisms; or d) stress response mechanisms are more general than previous thought and predator-induced stress produces similar physiological responses as temperature-induced stress.


Apart from inducing morphology changes, predator cues promoted an increase in CTmax and optimum temperature of Hyla versicolor tadpoles. As such, in the presence of predators, we can expect tadpoles to have greater warming tolerance and broader thermal safety margins. These changes might indirectly help tadpoles cope with increasing environmental temperatures. The herbicide Roundup is not only toxic to amphibians (and lethal over certain concentrations), but it also produces changes in morphology [17]. With this work, we now know that it also interferes, to some extent, with the thermal physiology of tadpoles (in particular in the thermal performance curves), although the effect on warming tolerance and thermal safety margins appears to be marginal. However, Roundup is just one of hundreds of chemicals currently used in anthropogenic activities (e.g., agriculture) and tadpoles can face predation by a wide variety of predator species. Because combinations of pesticides, which are a common situation in natural environments, can have greater impacts than each pesticide alone [79], future studies should test whether combinations of pesticides and predators could have different effects on the thermal physiology of organisms.

In the current scenario of climate change, it is important that we understand the physiological mechanisms underlying tolerance to abiotic stress [80], [81] and the sensitivity of organisms to changes in the environment [80], [82]. However, it also is important that we understand the indirect effects of physiological responses (in particular thermal physiology) on species interactions, such as predation, competition and disease transmission [2]. Therefore, understanding the plasticity of thermal performance curves and thermal limits (CTmax and CTmin) and how these parameters are altered by environmental stressors may be critical to understanding how physiological variation can influence a species' response to climate change [83].

Supporting Information

Methods S1.

Appendices 1–4. Appendix 1, Detailed information on laboratory conditions for rearing tadpoles during acclimation for the experiments. Appendix 2, Description of method and apparatus used for measuring CTmax. Appendix 3, TMV method equation for calculating thermal performance curve's parameters. Appendix 4, Description of the side-view landmarks and semi-landmarks, and linear measurements in a hypothetical tadpole.


Raw Data S1.

CTmax and swimming performance database.


TPS Data S1.

Geometric morphometry database.



Our thanks to Aaron Stoler, Jessica Hua, Will Brogan, R.J. Bendis, Keri Simonette, Zach Zbinden, Lindsey Freed, Patrick Noyes, Beverly French, Chris Hensley, Elena Valdés, Sara Tripodi and Casilda Maldonado for their assistance with the experiment. We also thank the anonymous reviewers for comments on previous versions of this manuscript.

Author Contributions

Conceived and designed the experiments: MK JH RAR. Performed the experiments: MK JH. Analyzed the data: MK CC. Contributed reagents/materials/analysis tools: MK MT RAR. Wrote the paper: MK JH HD MT CC RAR.


  1. 1. Kearney M, Porter W (2009) Mechanistic niche modelling: combining physiological and spatial data to predict species' ranges. Ecol Lett 12: 334–350.
  2. 2. Helmuth B, Broitman BR, Yamane L, Gilman SE, Mach K, et al. (2010) Organismal climatology: analyzing environmental variability at scales relevant to physiological stress. J Exp Biol 213: 995–1003.
  3. 3. Chown SL, Terblanche JS (2007) Physiological diversity in insects: ecological and evolutionary contexts. Adv In Insect Phys 33: 50–152.
  4. 4. Pörtner HO, Farrell AP (2008) Physiology and climate change. Science 322: 690–692.
  5. 5. Hoffmann AA (2010) Physiological climatic limits in Drosophila: patterns and implications. J Exp Biol 213: 870–880.
  6. 6. Chown SL, Hoffmann AA, Kristensen TN, Angilletta Jr MJ, Stenseth NC, et al. (2010) Adapting to climate change: a perspective from evolutionary physiology. Clim Res 43: 3–15.
  7. 7. Ghalambor CK, McKay JK, Carroll SP, Reznick DN (2007) Adaptive versus non-adaptive phenotypic plasticity and the potential for contemporary adaptation in new environments. Funct Ecol 21: 394–407.
  8. 8. Arnqvist G, Johansson F (1998) Ontogenic reaction norms of predator- induced defensive morphology in dragonfly larvae. Ecology 79: 1847–1858.
  9. 9. Relyea RA (2002) The many faces of predation: how induction, selection, and thinning combine to alter prey phenotypes. Ecology 83: 1953–1964.
  10. 10. Domenici P, Turesson H, Brodersen J, Brönmark C (2008) Predator-induced morphology enhances escape locomotion in crucian carp. Proc R Soc B Biol Sci 275: 195–201.
  11. 11. Brönmark C, Lakowitz T, Hollander J (2011) Predator-induced morphological plasticity across local populations of a freshwater snail. PLoS One 6: e21773.
  12. 12. Poupardin R, Reynaud S, Strode C, Ranson H, Vontas J, et al. (2008) Cross-induction of detoxification genes by environmental xenobiotics and insecticides in the mosquito Aedes aegypti: impact on larval tolerance to chemical insecticides. Insect Biochem Mol Biol 38: 540–551.
  13. 13. Hua J, Morehouse N, Relyea RA (2013) Pesticide resistance in amphibians: Induced resistance in susceptible populations, constitutive tolerance in tolerant populations. Evol Appl 6: 1028–1040.
  14. 14. Hanazato T (1991) Pesticides as chemical agents inducing helmet formation in Daphnia ambigua. Freshw Biol 26: 419–424.
  15. 15. Barry MJ (1998) Endosulfan-enhanced crest induction in Daphnia longicephala: evidence for cholinergic innervation of kairomone receptors. J Plankton Res 20: 1219–1231.
  16. 16. Oda S, Kato Y, Watanabe H, Tatarazako N, Iguchi T (2011) Morphological changes in Daphnia galeata induced by a crustacean terpenoid hormone and its analog. Environ Toxicol Chem 30: 232–238.
  17. 17. Relyea RA (2012) New effects of Roundup on amphibians: predators reduce herbicide mortality; herbicides induce antipredator morphology. Ecol Appl 22: 634–647.
  18. 18. Hanazato T (1999) Anthropogenic chemicals (insecticides) disturb natural organic chemical communication in the plankton community. Environ Pollut 105: 137–142.
  19. 19. Barry MJ (1999) The effects of a pesticide on inducible phenotypic plasticity in Daphnia. Environ Pollut 104: 217–224.
  20. 20. Barry MJ (2000) Effects of endosulfan on Chaoborus-induced life-history shifts and morphological defenses in Daphnia pulex. J Plankton Res 22: 1705–1718.
  21. 21. Sakamoto M, Chang K-H, Hanazato T (2006) Inhibition of development of anti-predator morphology in the small cladoceran Bosmina by an insecticide: impact of an anthropogenic chemical on prey-predator interactions. Freshw Biol 51: 1974–1983.
  22. 22. Hutchison VH (1961) Critical thermal maxima in salamanders. Physiol Zool 34: 92–125.
  23. 23. Lutterschmidt WI, Hutchison VH (1997) The critical thermal maximum: history and critique. Can J Zool 75: 1561–1574.
  24. 24. Heath AG, Cech JJ, Brink L, Moberg P, Zinkl JG (1997) Physiological responses of fathead minnow larvae to rice pesticides. Ecotoxicol Environ Saf 37: 280–288.
  25. 25. Angilletta Jr MJ (2009) Thermal Adaptation: A Theoretical and Empirical Synthesis. Oxford: Oxford University Press, USA. 320 p.
  26. 26. Van Buskirk J (2002) A comparative test of the adaptive plasticity hypothesis: relationships between habitat and phenotype in anuran larvae. Am Nat 160: 87–102.
  27. 27. Miner BG, Sultan SE, Morgan SG, Padilla DK, Relyea RA (2005) Ecological consequences of phenotypic plasticity. Trends Ecol Evol 20: 685–692.
  28. 28. Relyea RA (2007) Getting out alive: how predators affect the decision to metamorphose. Oecologia 152: 389–400.
  29. 29. Hayes TB, Falso P, Gallipeau S, Stice M (2010) The cause of global amphibian declines: a developmental endocrinologist's perspective. J Exp Biol 213: 921–933.
  30. 30. Wagner N, Reichenbecher W, Teichmann H, Tappeser B, Lötters S (2013) Questions concerning the potential impact of glyphosate-based herbicides on amphibians. Environ Toxicol Chem 32: 1688–1700.
  31. 31. Huey RB, Kingsolver JG (1989) Evolution of thermal sensitivity of ectotherm performance. Trends Ecol Evol 4: 131–135.
  32. 32. Van Buskirk J, McCollum SA (2000) Influence of tail shape on tadpole swimming performance. J Exp Biol 203: 2149–2158.
  33. 33. Hopkins WA, Winne CT, DuRant SE (2005) Differential swimming performance of two natricine snakes exposed to a cholinesterase-inhibiting pesticide. Environ Pollut 133: 531–540.
  34. 34. Relyea RA (2005) The impact of insecticides and herbicides on the biodiversity and productivity of aquatic communities. Ecol Appl 15: 618–627.
  35. 35. Jones DK, Hammond JI, Relyea RA (2011) Competitive stress can make the herbicide Roundup more deadly to larval amphibians. Environ Toxicol Chem 30: 446–454.
  36. 36. Petranka JW, Kats LB, Sih A (1987) Predator-prey interactions among fish and larval amphibians: use of chemical cues to detect predatory fish. Anim Behav 35: 420–425.
  37. 37. Kats LB, Petranka JW, Sih A (1988) Antipredator defenses and the persistence of amphibian larvae with fishes. Ecology 69: 1865–1870.
  38. 38. Relyea RA, Auld JR (2005) Predator- and competitor-induced plasticity: how changes in foraging morphology affect phenotypic trade-offs. Ecology 86: 1723–1729.
  39. 39. Jones DK, Hammond JI, Relyea RA (2010) Roundup and amphibians: the importance of concentration, application time, and stratification. Environ Toxicol Chem 29: 2016–2025.
  40. 40. Brattstrom BH (1968) Thermal acclimation in anuran amphibians as a function of latitude and altitude. Comp Biochem Physiol 24: 93–111.
  41. 41. Relyea RA (2003) Predators come and predators go: the reversibility of predator-induced traits. Ecology 84: 1840–1848.
  42. 42. Gosner KL (1960) A simplified table for staging anuran embryos and larvae with notes on identification. Herpetologica 16: 183–190.
  43. 43. Sherman E (1980) Ontogenetic change in thermal tolerance of the toad Bufo woodhousii fowleri. Comp Biochem Physiol Part A 65A: 227–230.
  44. 44. Terblanche JS, Hoffmann AA, Mitchell KA, Rako L, le Roux PC, et al. (2011) Ecologically relevant measures of tolerance to potentially lethal temperatures. J Exp Biol 214: 3713–3725.
  45. 45. Gvoždík L, Van Damme R (2006) Triturus newts defy the running-swimming dilemma. Evolution 60: 2110–2121.
  46. 46. Gvoždík L, Van Damme R (2008) The evolution of thermal performance curves in semi-aquatic newts: Thermal specialists on land and thermal generalists in water? J Therm Biol 33: 395–403.
  47. 47. Learning in Motion, Inc (2004) Measurement in Motion software, v3.0.
  48. 48. Arendt JD (2009) Influence of sprint speed and body size on predator avoidance in New Mexican spadefoot toads (Spea multiplicata). Oecologia 159: 455–461.
  49. 49. Arendt J (2010) Morphological correlates of sprint swimming speed in five species of spadefoot toad tadpoles: Comparison of morphometric methods. J Morphol 271: 1044–1052.
  50. 50. Van Damme R, Van Dooren TJ (1999) Absolute versus per unit body length speed of prey as an estimator of vulnerability to predation. Anim Behav 57: 347–352.
  51. 51. Izem R, Kingsolver JG (2005) Variation in continuous reaction norms: quantifying directions of biological interest. Am Nat 166: 277–289.
  52. 52. David JR, Gibert P, Gravot E, Petavy G, Morin J-P, et al. (1997) Phenotypic plasticity and developmental temperature in Drosophila: analysis and significance of reaction norms of morphometrical traits. J Therm Biol 22: 441–451.
  53. 53. Huey RB, Stevenson RD (1979) Integrating thermal physiology and ecology of ectotherms: a discussion of approaches. Am Zool 366: 357–366.
  54. 54. The MathWorks, Inc.. (2009) MATLAB v7.8 and Statistics Toolbox Release 2009a. Natick, Massachusetts, United States.
  55. 55. Dayton GH, Saenz D, Baum KA, Langerhans RB, DeWitt TJ (2005) Body shape, burst speed and escape behavior of larval anurans. Oikos 111: 582–591.
  56. 56. Rohlf FJ (2010) tpsDig2, digitize landmarks and outlines, version 2.16. Department of Ecology and Evolution, State University of New York at Stony Brook.
  57. 57. Rohlf FJ (2010) tpsRelw, relative warps analysis, version 1.49. Department of Ecology and Evolution, State University of New York at Stony Brook.
  58. 58. Bookstein FL (1991) Morphometric tools for landmark data. Cambridge/New York/Port Chester/Melbourne/Sydney: Cambridge University Press.
  59. 59. Klingenberg CP (2011) MorphoJ: an integrated software package for geometric morphometrics. Mol Ecol Resour 11: 353–357.
  60. 60. Patra RW, Chapman JC, Lim RP, Gehrke PC (2007) The effects of three organic chemicals on the upper thermal tolerances of four freshwater fishes. Environ Toxicol Chem 26: 1454–1459.
  61. 61. Carrier R, Beitinger TL (1988) Reduction in thermal tolerance of Notropis lutrensis and Pimephales promelas exposed to cadmium. Water Res 22: 511–515.
  62. 62. Lydy MJ, Wissing TE (1988) Effect of sublethal concentrations of copper on the critical thermal maxima (CTMax) of the fantail (Etheostoma flabellare) and johnny (E. nigrum) darters. Aquat Toxicol 12: 311–321.
  63. 63. Relyea RA, Jones DK (2009) The toxicity of Roundup Original Max to 13 species of larval amphibians. Environ Toxicol Chem 28: 2004–2008.
  64. 64. Navas CA, Gomes FR, Carvalho JE (2008) Thermal relationships and exercise physiology in anuran amphibians: integration and evolutionary implications. Comp Biochem Physiol Part A 151: 344–362.
  65. 65. Terblanche JS, Deere JA, Clusella-Trullas S, Janion C, Chown SL (2007) Critical thermal limits depend on methodological context. Proc R Soc London B 274: 2935–2942.
  66. 66. Chown SL, Jumbam KR, Sørensen JG, Terblanche JS (2009) Phenotypic variance, plasticity and heritability estimates of critical thermal limits depend on methodological context. Funct Ecol 23: 133–140.
  67. 67. Mitchell KA, Hoffmann AA (2010) Thermal ramping rate influences evolutionary potential and species differences for upper thermal limits in Drosophila. Funct Ecol 24: 694–700.
  68. 68. Rezende EL, Tejedo M, Santos M (2011) Estimating the adaptive potential of critical thermal limits: methodological problems and evolutionary implications. Funct Ecol 25: 111–121.
  69. 69. Mahoney JJ, Hutchison VH (1969) Photoperiod acclimation and 24-hour variations in the critical thermal maxima of a tropical and a temperate frog. Oecologia 2: 143–161.
  70. 70. Deutsch CA, Tewksbury JJ, Huey RB, Sheldon KS, Ghalambor CK, et al. (2008) Impacts of climate warming on terrestrial ectotherms across latitude. Proc Natl Acad Sci 105: 6668–6672.
  71. 71. Duarte H, Tejedo M, Katzenberger M, Marangoni F, Baldo D, et al. (2012) Can amphibians take the heat? Vulnerability to climate warming in subtropical and temperate larval amphibian communities. Glob Chang Biol 18: 412–421.
  72. 72. Smith JB, Schneider SH, Oppenheimer M, Yohe GW, Hare W, et al. (2009) Assessing dangerous climate change through an update of the Intergovernmental Panel on Climate Change (IPCC) “reasons for concern.”. Proc Natl Acad Sci 106: 4133.
  73. 73. Kingsolver JG, Huey RB (1998) Evolutionary analyses of morphological and physiological plasticity in thermally variable environments. Am Zool 545: 1–16.
  74. 74. Condon CH, Chenoweth SF, Wilson RS (2010) Zebrafish take their cue from temperature but not photoperiod for the seasonal plasticity of thermal performance. J Exp Biol 213: 3705–3709.
  75. 75. Lachenicht MW, Clusella-Trullas S, Boardman L, Le Roux C, Terblanche JS (2010) Effects of acclimation temperature on thermal tolerance, locomotion performance and respiratory metabolism in Acheta domesticus L. (Orthoptera: Gryllidae). J Insect Physiol 56: 822–830.
  76. 76. Gvoždík L, Puky M, Šugerková M (2007) Acclimation is beneficial at extreme test temperatures in the Danube crested newt, Triturus dobrogicus (Caudata, Salamandridae). Biol J Linn Soc 90: 627–636.
  77. 77. Relyea RA (2001) Morphological and behavioral plasticity of larval anurans in response to different predators. Ecology 82: 523–540.
  78. 78. Glennemeier KA, Denver RJ (2002) Small changes in whole-body corticosterone content affect larval Rana pipiens fitness components. Gen Comp Endocrinol 127: 16–25.
  79. 79. Relyea RA (2009) A cocktail of contaminants: how mixtures of pesticides at low concentrations affect aquatic communities. Oecologia 159: 363–376.
  80. 80. Gilman SE, Wethey DS, Helmuth B (2006) Variation in the sensitivity of organismal body temperature to climate change over local and geographic scales. Proc Natl Acad Sci 103: 9560–9565.
  81. 81. Tewksbury JJ, Huey RB, Deutsch CA (2008) Putting the Heat on Tropical Animals. Science (80-) 320: 1296–1297.
  82. 82. Porter WP, Gates DM (1969) Thermodynamic equilibria of animals with environment. Ecol Monogr 39: 227–244.
  83. 83. Buckley LB (2008) Linking traits to energetics and population dynamics to predict lizard ranges in changing environments. Am Nat 171: E1–E19.