Long-Term Resilience of Primary Sex Ratios in a Species with Temperature-Dependent Sex Determination after Decades of Climate Warming.

AbstractSpecies with environmental sex determination (ESD) have persisted through deep time, despite massive environmental perturbation in the geological record. Understanding how species with temperature-dependent sex determination (TSD), a type of ESD, persist through climate change is particularly timely given the current climate crisis, as highly biased sex ratios and extinction are predicted. Since 1982, we have studied primary sex ratios of a reptile with TSD (Chelydra serpentina). Primary sex ratios remained unchanged over time, despite warming in the environment. Resilience of the primary sex ratio occurred via a portfolio effect, realized through remarkable intra-annual variation in nest-level sex ratios, leading to a relatively consistent mean annual sex ratio. Intra-annual variation in nest-level sex ratios was related to variation in egg burial depth coupled with large clutch sizes, creating thermal gradients in the nest and promoting mixed-sex clutches. Furthermore, both locally and globally, sustained increases in nighttime air temperature contribute more to warming than increases in daily maximum temperature, but development rate was affected more strongly by maximum daily air temperature, conferring additional resilience to overall warming. Our study suggests that some TSD species may be resilient to warming and provides an example of how ESD may persist under environmental change.

Climate-associated decline of body condition in a fossorial salamander.

Temperate ectotherms have responded to recent environmental change, likely due to the direct and indirect effects of temperature on key life cycle events. Yet, a substantial number of ectotherms are fossorial, spending the vast majority of their lives in subterranean microhabitats that are assumed to be buffered against environmental change. Here, we examine whether seasonal climatic conditions influence body condition (a measure of general health and vigor), reproductive output, and breeding phenology in a northern population of fossorial salamander (Spotted Salamander, Ambystoma maculatum). We found that breeding body condition declined over a 12-year monitoring period (2008-2019) with warmer summer and autumn temperatures at least partly responsible for the observed decline in body condition. Our findings are consistent with the hypothesis that elevated metabolism drives the negative association between temperature and condition. Population-level reproduction, assessed via egg mass counts, showed high interannual variation and was weakly influenced by autumn temperatures. Salamander breeding phenology was strongly correlated with lake ice melt but showed no long-term temporal trend (1986-2019). Climatic warming in the region, which has been and is forecasted to be strongest in the summer and autumn, is predicted to lead to a 5%-27% decline in salamander body condition under realistic near-future climate scenarios. Although the subterranean environment offers a thermal buffer, the observed decline in condition and relatively strong effect of summer temperature on body condition suggest that fossorial salamanders are sensitive to the effects of a warming climate. Given the diversity of fossorial taxa, heightened attention to the vulnerability of subterranean microhabitat refugia and their inhabitants is warranted amid global climatic change.

Extinction and the temporal distribution of macroevolutionary bursts.

Phenotypic evolution through deep time is slower than expected from microevolutionary rates. This is the paradox of stasis. Previous models suggest stasis occurs because populations track adaptive peaks that remain relatively stable on million-year intervals, raising the equally perplexing question of why these large changes are so rare. Here, we consider the possibility that peaks can move more rapidly than populations can adapt, resulting in extinction. We model peak movement with explicit population dynamics, parameterized with published microevolutionary estimates. Allowing extinction greatly increases the parameter space of peak movements that yield the appearance of stasis observed in real data through deep time. Extreme peak displacements, regardless of their frequency, will rarely result in an equivalent degree of trait evolution because of extinction. Thus, larger peak displacements will rarely be inferred using trait data from extant species or observed in fossil records. Our work highlights population ecology as an important contributor to macroevolutionary dynamics, presenting an alternative perspective on the paradox of stasis, where apparent constraint on phenotypic evolution in deep time reflects our restricted view of the subset of earth's lineages that were fortunate enough to reside on relatively stable peaks.

Maternal provisioning and fluctuating thermal regimes enhance immune response in a reptile with temperature-dependent sex determination.

The Charnov-Bull model of differential fitness is often used to explain the evolution and maintenance of temperature-dependent sex determination (TSD). Most tests of the model focus on morphological proxies of fitness, such as size traits, whereas early life physiological traits that are closely related to lifetime fitness might provide a framework for generalizing the Charnov-Bull model across taxa. One such trait is the strength of the early-life immune response, which is strongly linked to early-life survival and fitness. Here, we manipulated temperature, variance in temperature, and sex to test the Charnov-Bull model using a physiological trait, immune system strength, in the snapping turtle (Chelydra serpentina). We found no evidence of sex-specific differences in bactericidal capacity of hatchling blood, and no evidence that mean temperature influences bactericidal capacity. However, we did find that fluctuating incubation temperature (i.e. a more naturalized incubation regime) is associated with a greater bactericidal capacity compared with constant temperature incubation. We also found that egg mass, a proxy for maternal provisioning, is positively associated with bactericidal capacity. Our findings suggest that the evolution of temperature-dependent sex determination in reptiles is unrelated to our measure of early-life innate immunity. Our study also underlines how immune response is condition dependent in early life, and questions the biological relevance of constant temperature incubation in experimental studies on ectotherm development.

Adaptive Maternal Investment in the Wild? Links between Maternal Growth Trajectory and Offspring Size, Growth, and Survival in Contrasting Environments.

Life-history theory predicts that investment per offspring should correlate negatively with the quality of the environment that offspring are anticipated to encounter; parents may use their own experience as juveniles to predict this environment and may modulate offspring traits, such as growth capacity and initial size. We manipulated nutrient levels in the juvenile habitat of wild Atlantic salmon (Salmo salar) to investigate the hypothesis that the egg size that maximizes juvenile growth and survival depends on environmental quality. We also tested whether offspring traits were related to parental growth trajectory. Mothers that grew fast when young produced more offspring and smaller offspring than mothers that grew slowly to reach the same size. Despite their size disadvantage, offspring of faster-growing mothers grew faster than those of slower-growing mothers in all environments, counter to the expectation that they would be competitively disadvantaged. However, they had lower relative survival in environments where the density of older predatory/competitor fish was relatively high. These links between maternal (but not paternal) growth trajectory and offspring survival rate were independent of egg size, underscoring that mothers may be adjusting egg traits other than size to suit the environment their offspring are anticipated to face.

Measurement and modelling of primary sex ratios for species with temperature-dependent sex determination.

For many oviparous animals, incubation temperature influences sex through temperature-dependent sex determination (TSD). Although climate change may skew sex ratios in species with TSD, few available methods predict sex under natural conditions, fewer still are based on mechanistic hypotheses of development, and field tests of existing methods are rare. We propose a new approach that calculates the probability of masculinization (PM) in natural nests. This approach subsumes the mechanistic hypotheses describing the outcome of TSD, by integrating embryonic development with the temperature-dependent reaction norm for sex determination. Further, we modify a commonly used method of sex ratio estimation, the constant temperature equivalent (CTE), to provide quantitative estimates of sex ratios. We test our new approaches using snapping turtles (Chelydra serpentina). We experimentally manipulated nests in the field, and found that the PM method is better supported than the modified CTE, explaining 69% of the variation in sex ratios across 27 semi-natural nests. Next, we used the PM method to predict variation in sex ratios across 14 natural nests over 2 years, explaining 67% of the variation. We suggest that the PM approach is effective and broadly applicable to species with TSD, particularly for forecasting how sex ratios may respond to climate change. Interestingly, we also found that the modified CTE explained up to 64% of variation in sex ratios in a Type II TSD species, suggesting that our modifications will be useful for future research. Finally, our data suggest that the Algonquin Park population of snapping turtles possesses resilience to biased sex ratios under climate change.

Anthropogenic nest sites provide warmer incubation environments than natural nest sites in a population of oviparous reptiles near their northern range limit.

Oviposition site choice affects a host of offspring phenotypes and directly impacts maternal fitness. Recent evidence suggests that oviparous reptiles often select nest sites where the landscape has been altered by anthropogenic activity, whereas natural nest sites are less often used. We leverage a long-term study of snapping turtle (Chelydra serpentina) to identify natural nest sites and anthropogenic nest sites and to compare habitat variables among nest site types. Natural and anthropogenic nest sites did not differ in average canopy closure, distance to nearest water, substrate composition, or aspect. However, anthropogenic nest sites had less ground-level vegetation and greater soil brightness, and were 3.3 °C warmer than natural nests during incubation. We used the Schoolfield model of poikilotherm development to assess differences in development rate between natural and anthropogenic nests. Because of the difference in temperature, embryos in anthropogenic nests were predicted to have undergone nearly twice as much development as embryos in natural nests during incubation. We outline why the evolution of fast embryonic development rate cannot compensate indefinitely for the low temperature incubation regimes that become increasingly prevalent at northern range margins, thereby underlining why maternal nest site choice of relatively warm anthropogenic sites may help oviparous reptiles persist in thermally constrained environments. Future research should aim to quantify both the thermal benefits of anthropogenic nest sites, as well as associated fitness costs (e.g., increased adult mortality) to elucidate whether anthropogenic disturbance of the landscape can be an ecological trap or serve a net benefit to some reptiles in northern environments.

Oxygen Limitation at the Larval Stage and the Evolution of Maternal Investment per Offspring in Aquatic Environments.

Oxygen limitation and surface area to volume relationships of the egg were long thought to constrain egg size in aquatic environments, but more recent evidence indicates that egg size per se does not influence oxygen availability to embryos. Here, we suggest that investment per offspring is nevertheless constrained in aquatic anamniotes by virtue of oxygen transport in free-living larvae. Drawing on the well-supported assumption that oxygen limitation is relatively pronounced in aquatic versus terrestrial environments and that oxygen limitation is particularly severe in warm aquatic environments, we employ comparative methods in the Amphibia to investigate this problem. Across hundreds of species and two major amphibian clades, the slope of species mean egg diameter over habitat temperature is negative for species with aquatic larvae but is positive or neutral for species featuring terrestrial eggs and no larvae. Yet across species with aquatic larvae, the negative slope of egg diameter over temperature is similar whether eggs are laid terrestrially or aquatically, consistent with an oxygen constraint arising at the larval stage. Finally, egg size declines more strongly with temperature for species that cannot breathe aerially before metamorphosis compared with those that can. Our results suggest that oxygen transport in larvae (not eggs) constrains investment per offspring. This study further extends the generality of temperature-dependent oxygen limitation as a mechanism driving the temperature-size rule in aquatic systems.

A new method of estimating thermal performance of embryonic development rate yields accurate prediction of embryonic age in wild reptile nests.

Temperature has a strong effect on ectotherm development rate. It is therefore possible to construct predictive models of development that rely solely on temperature, which have applications in a range of biological fields. Here, we leverage a reference series of development stages for embryos of the turtle Chelydra serpentina, which was described at a constant temperature of 20 °C. The reference series acts to map each distinct developmental stage onto embryonic age (in days) at 20 °C. By extension, an embryo taken from any given incubation environment, once staged, can be assigned an equivalent age at 20 °C. We call this concept "Equivalent Development", as it maps the development stage of an embryo incubated at a given temperature to its equivalent age at a reference temperature. In the laboratory, we used the concept of Equivalent Development to estimate development rate of embryos of C. serpentina across a series of constant temperatures. Using these estimates of development rate, we created a thermal performance curve measured in units of Equivalent Development (TPCED). We then used the TPCED to predict developmental stage of embryos in several natural turtle nests across six years. We found that 85% of the variation of development stage in natural nests could be explained. Further, we compared the predictive accuracy of the model based on the TPCED to the predictive accuracy of a degree-day model, where development is assumed to be linearly related to temperature and the amount of accumulated heat is summed over time. Information theory suggested that the model based on the TPCED better describes variation in developmental stage in wild nests than the degree-day model. We suggest the concept of Equivalent Development has several strengths and can be broadly applied. In particular, studies on temperature-dependent sex determination may be facilitated by the concept of Equivalent Development, as development age maps directly onto the developmental series of the organism, allowing critical periods of sex determination to be delineated without invasive sampling, even under fluctuating temperature.

Phenotypic plasticity of nest timing in a post-glacial landscape: how do reptiles adapt to seasonal time constraints?

Life histories evolve in response to constraints on the time available for growth and development. Nesting date and its plasticity in response to spring temperature may therefore be important components of fitness in oviparous ectotherms near their northern range limit, as reproducing early provides more time for embryos to complete development before winter. We used data collected over several decades to compare air temperature and nest date plasticity in populations of painted turtles and snapping turtles from a relatively warm environment (southeastern Michigan) near the southern extent of the last glacial maximum to a relatively cool environment (central Ontario) near the northern extent of post-glacial recolonization. For painted turtles, population-level differences in reaction norm elevation for two phenological traits were consistent with adaptation to time constraints, but no differences in reaction norm slopes were observed. For snapping turtle populations, the difference in reaction norm elevation for a single phenological trait was in the opposite direction of what was expected under adaptation to time constraints, and no difference in reaction norm slope was observed. Finally, among-individual variation in individual plasticity for nesting date was detected only in the northern population of snapping turtles, suggesting that reaction norms are less canalized in this northern population. Overall, we observed evidence of phenological adaptation, and possibly maladaptation, to time constraints in long-lived reptiles. Where present, (mal)adaptation occurred by virtue of differences in reaction norm elevation, not reaction norm slope. Glacial history, generation time, and genetic constraint may all play an important role in the evolution of phenological timing and its plasticity in long-lived reptiles.

Risk assessment of inbreeding and outbreeding depression in a captive-breeding program.

Captive-breeding programs can be implemented to preserve the genetic diversity of endangered populations such that the controlled release of captive-bred individuals into the wild may promote recovery. A common difficulty, however, is that programs are founded with limited wild broodstock, and inbreeding can become increasingly difficult to avoid with successive generations in captivity. Program managers must choose between maintaining the genetic purity of populations, at the risk of inbreeding depression, or interbreeding populations, at the risk of outbreeding depression. We evaluate these relative risks in a captive-breeding program for 3 endangered populations of Atlantic salmon (Salmo salar). In each of 2 years, we released juvenile F(1) and F(2) interpopulation hybrids, backcrosses, as well as inbred and noninbred within-population crosstypes into 9 wild streams. Juvenile size and survival was quantified in each year. Few crosstype effects were observed, but interestingly, the relative fitness consequences of inbreeding and outbreeding varied from year to year. Temporal variation in environmental quality might have driven some of these annual differences, by exacerbating the importance of maternal effects on juvenile fitness in a year of low environmental quality and by affecting the severity of inbreeding depression differently in different years. Nonetheless, inbreeding was more consistently associated with a negative effect on fitness, whereas the consequences of outbreeding were less predictable. Considering the challenges associated with a sound risk assessment in the wild and given that the effect of inbreeding on fitness is relatively predictable, we suggest that risk can be weighted more strongly in terms of the probable outcome of outbreeding. Factors such as genetic similarities between populations and the number of generations in isolation can sometimes be used to assess outbreeding risk, in lieu of experimentation.

Environmental quality predicts optimal egg size in the wild.

Parents can maximize their reproductive success by balancing the trade-off between investment per offspring and fecundity. According to theory, environmental quality influences the relationship between investment per offspring and offspring fitness, such that well-provisioned offspring fare better when environmental quality is lower. A major prediction of classic theory, then, is that optimal investment per offspring will increase as environmental quality decreases. To test this prediction, we release over 30,000 juvenile Atlantic salmon (Salmo salar) into eight wild stream environments, and we monitor subsequent growth and survival of juveniles. We estimate the shape of the relationship between investment per offspring (egg size) and offspring fitness in each stream. We find that optimal egg size is greater when the quality of the stream environment is lower (as estimated by a composite index of habitat quality). Across streams, the mean size of stream gravel and the mean amount of incident sunlight are the most important individual predictors of optimal egg size. Within streams, juveniles recaptured in stream subsections that featured larger gravels and greater levels of sunlight also grew relatively quickly, an association that complements our cross-stream analyses. This study provides the first empirical verification that environmental quality alters the relationship between investment per offspring and offspring fitness, such that optimal investment per offspring increases as environmental quality decreases.

The relationship between offspring size and fitness: integrating theory and empiricism.

How parents divide the energy available for reproduction between size and number of offspring has a profound effect on parental reproductive success. Theory indicates that the relationship between offspring size and offspring fitness is of fundamental importance to the evolution of parental reproductive strategies: this relationship predicts the optimal division of resources between size and number of offspring, it describes the fitness consequences for parents that deviate from optimality, and its shape can predict the most viable type of investment strategy in a given environment (e.g., conservative vs. diversified bet-hedging). Many previous attempts to estimate this relationship and the corresponding value of optimal offspring size have been frustrated by a lack of integration between theory and empiricism. In the present study, we draw from C. Smith and S. Fretwell's classic model to explain how a sound estimate of the offspring size--fitness relationship can be derived with empirical data. We evaluate what measures of fitness can be used to model the offspring size--fitness curve and optimal size, as well as which statistical models should and should not be used to estimate offspring size--fitness relationships. To construct the fitness curve, we recommend that offspring fitness be measured as survival up to the age at which the instantaneous rate of offspring mortality becomes random with respect to initial investment. Parental fitness is then expressed in ecologically meaningful, theoretically defensible, and broadly comparable units: the number of offspring surviving to independence. Although logistic and asymptotic regression have been widely used to estimate offspring size-fitness relationships, the former provides relatively unreliable estimates of optimal size when offspring survival and sample sizes are low, and the latter is unreliable under all conditions. We recommend that the Weibull-1 model be used to estimate this curve because it provides modest improvements in prediction accuracy under experimentally relevant conditions.

Recurrent violations of invariant rules for offspring size: evidence from turtles and the implications for small clutch size models.

Smith and Fretwell's classic model predicts that parents can maximize fitness by dividing the energy available for reproduction into offspring of an optimal size. However, this model breaks down when clutch size is small (~1-10 offspring). Invariant rules are an extension of the Smith-Fretwell model, and these rules predict how offspring size will vary among and within individuals that produce small clutch sizes. Here, we provide a narrow test of invariant rules using three turtle species, then we synthesize and re-analyze existing data from 18 different species (comprising five Orders) to evaluate whether invariant rules are followed across broad taxa. We do not find support for most invariant rules in turtles, and our re-analysis demonstrates a general mismatch between observed and expected values across all taxa evaluated, suggesting that invariant rules fail to predict reproductive patterns in nature. Morphological constraints on offspring size and reproductive effort may be important reasons for disparities between theory and observation both in turtles and other taxa. Paradoxically, morphological constraints are most common in small-bodied species and individuals, but these same candidates are also those which produce the small clutch sizes that are necessary to test invariant rules, such that a fair test of invariant rules will often be challenging. Mismatches between theory and observation might also occur because theory assumes that mothers exert control over resource allocation to offspring. In fact, there is evidence of widespread genetic correlations among investment per offspring and reproductive effort, such that these traits are not independent.

Invasive Bullfrogs Maintain MHC Polymorphism Including Alleles Associated with Chytrid Fungal Infection.

Maintenance of genetic diversity at adaptive loci may facilitate invasions by non-native species by allowing populations to adapt to novel environments, despite the loss of diversity at neutral loci that typically occurs during founder events. To evaluate this prediction, we compared genetic diversity at major histocompatibility complex (MHC) and cytochrome b (cytb) loci from 20 populations of the American bullfrog (Rana catesbeiana) across theinvasive and native ranges in North America and quantified the presence of the pathogen Batrachochytrium dendrobatidis (Bd). Compared to native populations, invasive populations had significantly higher Bd prevalence and intensity, significantly higher pairwise MHC and cytb FST, and significantly lower cytb diversity, but maintained similar levels of MHC diversity. The two most common MHC alleles (LiCA_B and Rapi_33) were associated with a significant decreased risk of Bd infection, and we detected positive selection acting on four peptide binding residues. Phylogenetic analysis suggested invasive populations likely arose from a single founding population in the American Midwest with a possible subsequent invasion in the northwest. Overall, our study suggests that the maintenance of diversity at adaptive loci may contribute to invasion success and highlights the importance of quantifying diversity at functional loci to assess the evolutionary potential of invasive populations.

Body size-specific maternal effects on the offspring environment shape juvenile phenotypes in Atlantic salmon.

Positive associations between maternal investment per offspring and maternal body size have been explained as adaptive responses by females to predictable, body size-specific maternal influences on the offspring's environment. As a larger per-offspring investment increases maternal fitness when the quality of the offspring environment is low, optimal egg size may increase with maternal body size if larger mothers create relatively poor environments for their eggs or offspring. Here, we manipulate egg size and rearing environments (gravel size, nest depth) of Atlantic salmon (Salmo salar) in a 2 × 2 × 2 factorial experiment. We find that the incubation environment typical of large and small mothers can exert predictable effects on offspring phenotypes, but the nature of these effects provides little support to the prediction that smaller eggs are better suited to nest environments created by smaller females (and vice versa). Our data indicate that the magnitude and direction of phenotypic differences between small and large offspring vary among maternal nest environments, underscoring the point that removal of offspring from the environmental context in which they are provisioned in the wild can bias experimentally derived associations between offspring size and metrics of offspring fitness. The present study also contributes to a growing literature which suggests that the fitness consequences of egg size variation are often more pronounced during the early juvenile stage, as opposed to the egg or larval stage.

A simple model for the evolution of temperature-dependent sex determination explains the temperature sensitivity of embryonic mortality in imperiled reptiles.

A common reptile conservation strategy involves artificial incubation of embryos and release of hatchlings or juveniles into wild populations. Temperature-dependent sex determination (TSD) occurs in most chelonians, permitting conservation managers to bias sex ratios towards females by incubating embryos at high temperatures, ultimately allowing the introduction of more egg-bearing individuals into populations. Here, we revisit classic sex allocation theory and hypothesize that TSD evolved in some reptile groups (specifically, chelonians and crocodilians) because male fitness is more sensitive to condition (general health, vigor) than female fitness. It follows that males benefit more than females from incubation environments that confer high-quality phenotypes, and hence high-condition individuals. We predict that female-producing temperatures, which comprise relatively high incubation temperatures in chelonians and crocodilians, are relatively stressful for embryos and subsequent life stages. We synthesize data from 28 studies to investigate how constant temperature incubation affects embryonic mortality in chelonians with TSD. We find several lines of evidence suggesting that warm, female-producing temperatures are more stressful than cool, male-producing temperatures. Further, we find some evidence that pivotal temperatures (TPiv, the temperature that produces a 1:1 sex ratio) may exhibit a correlated evolution with embryonic thermal tolerance. If patterns of temperature-sensitive embryonic mortality are also indicative of chronic thermal stress that occurs post-hatching, then conservation programs may benefit from incubating eggs close to species-specific TPivs, thus avoiding high-temperature incubation. Indeed, our models predict that, on average, a sex ratio of >75% females can generally be achieved by incubating eggs only 1°C above TPiv. Of equal importance, we provide insight into the enigmatic evolution of TSD in chelonians, by providing support to the hypothesis that TSD evolution is related to the quality of the phenotype conferred by incubation temperature, with males produced in high-quality incubation environments.

The thermal ecology and physiology of reptiles and amphibians: A user's guide.

Research on the thermal ecology and physiology of free-living organisms is accelerating as scientists and managers recognize the urgency of the global biodiversity crisis brought on by climate change. As ectotherms, temperature fundamentally affects most aspects of the lives of amphibians and reptiles, making them excellent models for studying how animals are impacted by changing temperatures. As research on this group of organisms accelerates, it is essential to maintain consistent and optimal methodology so that results can be compared across groups and over time. This review addresses the utility of reptiles and amphibians as model organisms for thermal studies by reviewing the best practices for research on their thermal ecology and physiology, and by highlighting key studies that have advanced the field with new and improved methods. We end by presenting several areas where reptiles and amphibians show great promise for further advancing our understanding of how temperature relations between organisms and their environments are impacted by global climate change.

Density-dependent offspring interactions do not explain macroevolutionary scaling of adult size and offspring size.

Most life forms exhibit a correlated evolution of adult size (AS) and size at independence (SI), giving rise to AS-SI scaling relationships. Theory suggests that scaling arises because relatively large adults have relatively high reproductive output, resulting in strong density-dependent competition in early life, where large size at independence provides a competitive advantage to juveniles. The primary goal of our study is to test this density hypothesis, using large datasets that span the vertebrate tree of life (fishes, amphibians, reptiles, birds, and mammals). Our secondary goal is to motivate new hypotheses for AS-SI scaling by exploring how subtle variation in life-histories among closely related species is associated with variation in scaling. Our phylogenetically informed comparisons do not support the density hypothesis. Instead, exploration of AS-SI scaling among life-history variants suggests that steeper AS-SI scaling slopes are associated with evolutionary increases in size at independence. We suggest that a positive association between size at independence and juvenile growth rate may represent an important mechanism underlying AS-SI scaling, a mechanism that has been underappreciated by theorists. If faster juvenile growth is a consequence of evolutionary increases in size at independence, this may help offset the cost of delayed maturation, leading to steeper AS-SI scaling slopes.

Temperature-dependent oxygen limitation and the rise of Bergmann's rule in species with aquatic respiration.

Bergmann's rule is the propensity for species-mean body size to decrease with increasing temperature. Temperature-dependent oxygen limitation has been hypothesized to help drive temperature-size relationships among ectotherms, including Bergmann's rule, where organisms reduce body size under warm oxygen-limited conditions, thereby maintaining aerobic scope. Temperature-dependent oxygen limitation should be most pronounced among aquatic ectotherms that cannot breathe aerially, as oxygen solubility in water decreases with increasing temperature. We use phylogenetically explicit analyses to show that species-mean adult size of aquatic salamanders with branchial or cutaneous oxygen uptake becomes small in warm environments and large in cool environments, whereas body size of aquatic species with lungs (i.e., that respire aerially), as well as size of semiaquatic and terrestrial species do not decrease with temperature. We argue that oxygen limitation drives the evolution of small size in warm aquatic environments for species with aquatic respiration. More broadly, the stronger decline in size with temperature observed in aquatic versus terrestrial salamander species mirrors the relatively strong plastic declines in size observed previously among aquatic versus terrestrial invertebrates, suggesting that temperature-dependent oxygen availability can help drive patterns of plasticity, micro- and macroevolution.