Effects of heat and pesticide stress on life history, physiology and the gut microbiome of two congeneric damselflies that differ in stressor tolerance.

The combined impact of toxicants and warming on organisms is getting increased attention in ecotoxicology, but is still hard to predict, especially with regard to heat waves. Recent studies suggested that the gut microbiome may provide mechanistic insights into the single and combined stressor effects on their host. We therefore investigated effects of sequential exposure to a heat spike and a pesticide on both the phenotype (life history and physiology) and the gut microbiome composition of damselfly larvae. We compared the fast-paced Ischnura pumilio, which is more tolerant to both stressors, with the slow-paced I. elegans, to obtain mechanistic insights into species-specific stressor effects. The two species differed in gut microbiome composition, potentially contributing to their pace-of-life differences. Intriguingly, there was a general resemblance between the stressor response patterns in the phenotype and in the gut microbiome, whereby both species responded broadly similar to the single and combined stressors. The heat spike negatively affected the life history of both species (increased mortality, reduced growth rate), which could be explained not only by shared negative effects on physiology (inhibition of acetylcholinesterase, increase of malondialdehyde), but also by shared effects on gut bacterial species' abundances. The pesticide only had negative effects (reduced growth rate, reduced net energy budget) in I. elegans. The pesticide generated shifts in the bacterial community composition (e.g. increased abundance of Sphaerotilus and Enterobacteriaceae in the gut microbiome of I. pumilio), which potentially contributed to the relatively higher pesticide tolerance of I. pumilio. Moreover, in line with the response patterns in the host phenotype, the effects of the heat spike and the pesticide on the gut microbiome were mainly additive. By contrasting two species differing in stress tolerance, our results suggest that response patterns in the gut microbiome may improve our mechanistic understanding of single and combined stressor effects.

Rapid evolution of unimodal but not of linear thermal performance curves in Daphnia magna.

Species may cope with warming through both rapid evolutionary and plastic responses. While thermal performance curves (TPCs), reflecting thermal plasticity, are considered powerful tools to understand the impact of warming on ectotherms, their rapid evolution has been rarely studied for multiple traits. We capitalized on a 2-year experimental evolution trial in outdoor mesocosms that were kept at ambient temperatures or heated 4°C above ambient, by testing in a follow-up common-garden experiment, for rapid evolution of the TPCs for multiple key traits of the water flea Daphnia magna. The heat-selected Daphnia showed evolutionary shifts of the unimodal TPCs for survival, fecundity at first clutch and intrinsic population growth rate toward higher optimum temperatures, and a less pronounced downward curvature indicating a better ability to keep fitness high across a range of high temperatures. We detected no evolution of the linear TPCs for somatic growth, mass and development rate, and for the traits related to energy gain (ingestion rate) and costs (metabolic rate). As a result, also the relative thermal slope of energy gain versus energy costs did not vary. These results suggest the overall (rather than per capita) top-down impact of D. magna may increase under rapid thermal evolution.

Thermal and latitudinal patterns in pace-of-life traits are partly mediated by the gut microbiome.

The integration of life-history, physiological and behavioural traits into the pace-of-life generates a powerful framework to understand trait variation in nature both along environmental gradients and in response to environmental stressors. While the gut microbiome has been hypothesized as a candidate mechanism to underlie differentiation in the pace-of-life, this has been rarely studied. We investigated the role of the gut microbiome in contributing to the differentiation in pace-of-life and in thermal adaptation between populations of Ischnura elegans damselfly larvae inhabiting warmer low latitudes and colder high latitudes. We carried out a common-garden experiment, whereby we manipulated the exposure of the damselfly larvae to two key global warming factors: 4 °C warming and a 30 °C heat wave. Comparing the bacterial composition of the food source and the bacterioplankton indicated that damselfly larvae differentially take up bacteria from the surrounding environment and have a resident and functionally relevant microbiome. The gut microbiome differed between larvae of both latitudes, and this was associated with the host's latitudinal differentiation in activity, a key pace-of-life trait. Under heat wave exposure, the gut microbial community composition of high-latitude larvae converged towards that of the low-latitude larvae, with an increase in bacteria that likely are important in providing energy to cope with the heat wave. This suggests an adaptive latitude-specific shift in the gut microbiota matching the better ability of low-latitude hosts to deal with heat extremes. In general, our study provides evidence for the gut microbiome contributing to latitudinal differentiation in both the pace-of-life and in heat adaptation in natural populations.

A fast pace-of-life is traded off against a high thermal performance.

The integration of life-history, behavioural and physiological traits into a 'pace-of-life syndrome' is a powerful concept in understanding trait variation in nature. Yet, mechanisms maintaining variation in 'pace-of-life' are not well understood. We tested whether decreased thermal performance is an energetic cost of a faster pace-of-life. We characterized the pace-of-life of larvae of the damselfly Ischnura elegans from high-latitude and low-latitude regions when reared at 20°C or 24°C in a common-garden experiment, and estimated thermal performance curves for a set of behavioural, physiological and performance traits. Our results confirm a faster pace-of-life (i.e. faster growth and metabolic rate, more active and bold behaviour) in the low-latitude and in warm-reared larvae, and reveal increased maximum performance, Rmax, but not thermal optimum Topt, in low-latitude larvae. Besides a clear pace-of-life syndrome integration at the individual level, larvae also aligned along a 'cold-hot' axis. Importantly, a faster pace-of-life correlated negatively with a high thermal performance (i.e. higher Topt for swimming speed, metabolic rate, activity and boldness), which was consistent across latitudes and rearing temperatures. This trade-off, potentially driven by the energetically costly maintenance of a fast pace-of-life, may be an alternative mechanism contributing to the maintenance of variation in pace-of-life within populations.

Thermal plasticity and evolution shape predator-prey interactions differently in clear and turbid water bodies.

Warming and eutrophication negatively affect freshwater ecosystems by modifying trophic interactions and increasing water turbidity. We need to consider their joint effects on predator-prey interactions and how these depend on the thermal evolution of both predator and prey. We quantified how 4°C warming and algae-induced turbidity (that integrates turbidity per se and increased food for zooplankton prey) affect functional response parameters and prey population parameters in a common-garden experiment. We did so for all combinations of high- and low-latitude predator (damselfly larvae) and prey (water fleas) populations to assess the potential impact of thermal evolution of predators and/or prey at a high latitude under warming using a space-for-time substitution. We then modelled effects on the system stability (i.e. tendency to oscillate) under different warming, turbidity and evolutionary scenarios. Warming and turbidity had little effect on the functional response parameters of high-latitude predators. In contrast, warming and turbidity reduced the handling times of low-latitude predators. Moreover, warming increased the search rates of low-latitude predators in clear water but instead decreased these in turbid water. Warming increased stability (i.e. prevented oscillations) in turbid water (except for the 'high-latitude predator and high-latitude prey' system), mainly by decreasing the prey's carrying capacity and partly also by decreasing search rates, while it did not affect stability in clear water. Algae-induced turbidity generally decreased stability, mainly by increasing the prey's carrying capacity and partly also by increasing search rates. This resembles findings that nutrient enrichment can reduce the stability of trophic systems. The expected stability of the high-latitude trophic system under warming was dependent on the turbidity level: our results suggest that thermal plasticity tends to destabilize the high-latitude trophic system under warming in clear water but not in turbid water, and that thermal evolution of the predator will stabilize the high-latitude system under warming in turbid water but less so in clear water. The extent to which thermal plasticity and evolution shape trophic system stability under warming may strongly differ between clear and turbid water bodies, with their contributions having a more stabilizing role in turbid water.

Cryptic eco-evolutionary feedback in the city: Urban evolution of prey dampens the effect of urban evolution of the predator.

Most research on eco-evolutionary feedbacks focuses on ecological consequences of evolution in a single species. This ignores the fact that evolution in response to a shared environmental factor in multiple species involved in interactions could alter the net cumulative effect of evolution on ecology. We empirically tested whether urbanization-driven evolution in a predator (nymphs of the damselfly Ischnura elegans) and its prey (the water flea Daphnia magna) jointly shape the outcome of predation under simulated heatwaves. Both interactors show genetic trait adaptation to urbanization, particularly to higher temperatures. We cross-exposed common-garden reared damselflies and Daphnia from replicated urban and rural populations, and quantified predation rates and functional response traits. Urban damselfly nymphs showed higher encounter and predation rates than rural damselflies when exposed to rural prey, but this difference disappeared when they preyed on urban Daphnia. This represents a case of a cryptic evo-to-eco feedback, where the evolution of one species dampens the effects of the evolution of another species on their interaction strength. The effects of evolution of each single species were strong: the scenario in which only the predator or prey was adapted to urbanization resulted in a c. 250% increase in encounter rate and a c. 25% increase in predation rate, compared to the rural predator-rural prey combination. Our results provide unique evidence for eco-evolutionary feedbacks in cities, and underscore the importance of a multi-species approach in eco-evolutionary dynamics research.

Lower bioenergetic costs but similar immune responsiveness under a heat wave in urban compared to rural damselflies.

There is mounting evidence that the widespread phenotypic changes in response to urbanization may reflect adaptations caused by rapid evolutionary processes driven by urban-related stressors. Compared to increased habitat fragmentation and pollution, adaptations towards another typical urban-related stressor, that is higher and longer lasting very high temperatures (heat waves), are much less studied. Notably, the sensitivities to heat waves of life-history traits and important fitness-related physiological traits such as immune responsiveness and bioenergetic variables (energy availability, energy consumption and their balance) have never been contrasted between urban and rural populations. By conducting a laboratory common-garden experiment, we compared effects of a simulated heat wave on life history (survival and growth rate), immune responsiveness and bioenergetic variables between three urban and three rural populations of the damselfly Coenagrion puella. Because energy-mediated trade-off patterns may only be detected under energetically costly manipulations, all larvae were immune-challenged by simulating ectoparasitism by water mites. As expected, the simulated heat wave caused negative effects on nearly all response variables. The immune responsiveness, on the other hand, increased under the heat wave, consistent with a trade-off pattern between immune function and growth, and this similarly between urban and rural populations. A key finding was that urban larvae suffered less from the simulated heat wave compared to the rural larvae in terms of a lower heat wave-induced depletion in energy availability. This suggests an adaptation of urban populations to better cope with the stronger and more frequent heat waves in cities. Notably, this urbanization-driven evolution in the bioenergetic variables was not apparent in the absence of a heat wave. Given that changes in energy budgets have strong fitness consequences, our findings suggest that the evolved higher ability to cope with heat waves is fundamental for the survival of urban damselfly populations.

Higher mean and fluctuating temperatures jointly determine the impact of the pesticide chlorpyrifos on the growth rate and leaf consumption of a freshwater isopod.

There is growing evidence that both increases in mean temperature and the widespread daily temperature fluctuations (DTF) may increase pesticide toxicity. Nevertheless, the likely more stressful, realistic combination of the two warming-related stressors has rarely been considered in ecotoxicology. Moreover, we have little knowledge on whether these stressor combinations could impair ecosystem functioning. We examined the effect of the pesticide chlorpyrifos under an increased mean temperature (+4 °C, from 18 °C to 22 °C) and in the presence of DTF (constant and 8 °C) on two life-history traits (mortality and growth rate) and one ecologically important behavioural trait (feeding rate) in the freshwater isopod Asellus aquaticus. The chlorpyrifos concentration used, 0.2 µg/L, did not cause mortality in any thermal condition, nor did it cause sublethal effects at the mean temperature of 18 °C. A key finding was that growth rate was strongly reduced by the pesticide only under the combination of both a higher mean temperature and DTF, highlighting the importance of testing toxicity under this realistic thermal scenario. The leaf consumption of chlorpyrifos-exposed isopods increased at the higher mean temperature when this was kept constant, however, it lowered again towards control levels when DTF was induced, thereby contributing to the growth reduction at this most stressful condition. These alterations of growth and leaf degradation rates may impact nutrient recycling, a key ecosystem function. Our results highlight the importance of integrating both increases in mean temperature and in DTF to improve current and future ecological risk assessment of pesticides.

Strong species differences in life history do not predict oxidative stress physiology or sensitivity to an environmental oxidant.

Species typically align along a fast-slow life-history continuum, yet it is not clear to what extent oxidative stress physiology can be integrated with this continuum to form a 'pace-of-life syndrome', especially so in invertebrates. This is important, given the assumed role of oxidative stress in mediating life-history trade-offs, and the prediction that species with a faster pace should be more vulnerable to oxidative stress. We tested whether a species' life-history pace, here represented by its growth rate, can predict species-level differentiation in physiology and sensitivity to oxidative stress. Therefore, we exposed four species of Ischnura damselflies that strongly align along a fast-slow life-history continuum to different levels of ultraviolet (UV) radiation. We measured an extended set of physiological traits linked to the pace-of-life: standard metabolic rate, oxidative stress physiology (antioxidant enzymes and oxidative damage) and defence/condition traits (investment in immune function, energy storage and structural defence). Despite strong species differences in growth rate and physiology, growth rate did not predict species-level differentiation in physiology. Hence there was no support for the integration of metabolic rate, oxidative stress physiology or defence/condition traits into a species-level syndrome. Ultraviolet exposure affected nearly all traits: it reduced growth rate and increased metabolic rate, affected all oxidative stress physiology traits and increased the two defence traits (immune function, and melanin content). Nevertheless, the pace-of-life based on growth rate did not predict sensitivity to UV. Instead, the observed pattern of investment in structural UV defence (melanin) might have reduced the need for enzymatic antioxidant defence, this way potentially decoupling the covariation between the life-history pace and oxidative stress physiology. The absence of an integrated axis of life-history and physiological variation indicates no major constraints for the evolution of these traits among the studied damselfly species. Our study highlights that ecological differences between species may decouple covariation between species' life-history pace and their physiology, as well as their sensitivity to environmental stressors.

Latitude-associated evolution and drivers of thermal response curves in body stoichiometry.

Trait-based studies are needed to understand the plastic and genetic responses of organisms to warming. A neglected organismal trait is elemental composition, despite its potential to cascade into effects on the ecosystem level. Warming is predicted to shape elemental composition through shifts in storage molecules associated with responses in growth, body size and metabolic rate. Our goals were to quantify thermal response patterns in body composition and to obtain insights into their underlying drivers and their evolution across latitudes. We reconstructed the thermal response curves (TRCs) for body elemental composition [C (carbon), N (nitrogen) and the C:N ratio] of damselfly larvae from high- and low-latitude populations. Additionally, we quantified the TRCs for survival, growth and development rates and body size to assess local thermal adaptation, as well as the TRCs for metabolic rate and key macromolecules (proteins, fat, sugars and cuticular melanin and chitin) as these may underlie the elemental TRCs. All larvae died at 36°C. Up to 32°C, low-latitude larvae increased growth and development rates and did not suffer increased mortality. Instead, growth and development rates of high-latitude larvae were lower and levelled off at 24°C, and mortality increased at 32°C. This latitude-associated thermal adaptation pattern matched the 'hotter-is-better' hypothesis. With increasing temperatures, low-latitude larvae decreased C:N, while high-latitude larvae increased C:N. These patterns were driven by associated changes in N contents, while C contents did not respond to temperature. Consistent with the temperature-size rule and the thermal melanism hypothesis, body size and melanin levels decreased with warming. While all traits and associated macromolecules (except for metabolic rate that showed thermal compensation) assumed to underlie thermal responses in elemental composition showed thermal plasticity, these were largely independent and none could explain the stoichiometric TRCs. Our results highlight that thermal responses in elemental composition cannot be explained by traditionally assumed drivers, asking for a broader perspective including the thermal dependence of elemental fluxes. Another key implication is that thermal evolution can reverse the plastic stoichiometric thermal responses and hence reverse how warming may shape food web dynamics through changes in body composition at different latitudes.

Using natural laboratories to study evolution to global warming: contrasting altitudinal, latitudinal, and urbanization gradients.

Demonstrating the likelihood of evolution in response to global warming is important, yet challenging. We discuss how three spatial thermal gradients (latitudinal, altitudinal, and urbanization) can be used as natural laboratories to inform about the gradual thermal evolution of populations by applying a space-for-time substitution (SFTS) approach. We compare thermal variables and confounding non-thermal abiotic variables, methodological approaches and evolutionary aspects associated with each type of gradient. On the basis of an overview of recent insect studies, we show that a key assumption of SFTS, local thermal adaptation along these gradients, is often but not always met, requiring explicit validation. To increase realism when applying SFTS, we highlight the importance of integrating daily temperature fluctuations, multiple stressors and multiple interacting species. Finally, comparative studies, especially across gradient types, are important to provide more robust inferences of evolution under gradual global warming. Integrating these research directions will further strengthen the still underused, yet powerful SFTS approach to infer gradual evolution under global warming.

Evolution of geographic variation in thermal performance curves in the face of climate change and implications for biotic interactions.

We review the recent literature on geographic variation in insect thermal performance curves (TPCs). Despite strong thermal differences, there is often no change in TPCs across geographic gradients. When shifts occur, these are mostly vertical (indicating an overall shift in performance across temperatures, that is, countergradient or cogradient variation) and less horizontal (reflecting thermal adaptation). Based on this, using a space-for-time substitution approach, we generated likely evolutionary scenarios of TPC evolution to simulate the outcome of biotic interactions under future warming. We illustrate how taking evolution of the TPCs into account may strongly impact the predicted outcome of biotic interactions under climate warming. Importantly, both the type and the magnitude of the TPC shift was identified to be crucial to determine who will be winners and losers of biotic interactions.

Warming under seminatural outdoor conditions in the larval stage negatively affects insect flight performance.

Laboratory studies indicate global warming may cause changes in locomotor performance directly relevant for fitness and dispersal. Yet, this remains to be tested under seminatural settings, and the connection with warming-induced alterations in the underlying traits has been rarely studied. In an outdoor mesocosm experiment with the damselfly Ischnura elegans, 4°C warming in the larval stage decreased the flight muscle mass, which correlated with a lower flight endurance. Warming did not affect body mass, size or wing morphology. This illustrates how carry-over effects of warming under seminatural conditions during early development bridge metamorphosis and negatively impact locomotor performance through changes in a key flight-related trait.

Microgeographic differentiation in thermal performance curves between rural and urban populations of an aquatic insect.

The rapidly increasing rate of urbanization has a major impact on the ecology and evolution of species. While increased temperatures are a key aspect of urbanization ("urban heat islands"), we have very limited knowledge whether this generates differentiation in thermal responses between rural and urban populations. In a common garden experiment, we compared the thermal performance curves (TPCs) for growth rate and mortality in larvae of the damselfly Coenagrion puella from three urban and three rural populations. TPCs for growth rate shifted vertically, consistent with the faster-slower theoretical model whereby the cold-adapted rural larvae grew faster than the warm-adapted urban larvae across temperatures. In line with costs of rapid growth, rural larvae showed lower survival than urban larvae across temperatures. The relatively lower temperatures hence expected shorter growing seasons in rural populations compared to the populations in the urban heat islands likely impose stronger time constraints to reach a certain developmental stage before winter, thereby selecting for faster growth rates. In addition, higher predation rates at higher temperature may have contributed to the growth rate differences between urban and rural ponds. A faster-slower differentiation in TPCs may be a widespread pattern along the urbanization gradient. The observed microgeographic differentiation in TPCs supports the view that urbanization may drive life-history evolution. Moreover, because of the urban heat island effect, urban environments have the potential to aid in developing predictions on the impact of climate change on rural populations.

Carry-Over Effects Across Metamorphosis of a Pesticide on Female Lifetime Fitness Strongly Depend on Egg Hatching Phenology: A Longitudinal Study under Seminatural Conditions.

Current ecological risk assessment of pesticides fails to protect aquatic biodiversity. For the first time, we tested two potential reasons for this failure with regard to carry-over effects across metamorphosis: their dependence on hatching period, and the lack of studies quantifying adult fitness under seminatural conditions. Using the damselfly Coenagrion puella sampled from six populations, we designed an outdoor longitudinal one-year study starting from the egg stage. We exposed the aquatic larvae to the pesticide esfenvalerate (0.11 µg/L) during the initial microcosm part. Next, we monitored the lifetime fitness of the terrestrial adults in an insectary. Exposure to the pesticide negatively impacted not only larval traits, but also drastically reduced lifetime mating success of adult females. The impact of this postmetamorphic effect of the pesticide on the population level was three times more important than the effects in the larval stage. Importantly, this carry-over effect was only present in females that hatched early in the season, and was not mediated by metamorphic traits (age and mass at emergence). We provide proof-of-principle under seminatural conditions for two potential pitfalls that need to be considered when improving risk assessment: carry-over effects on adult fitness can (i) be much more important than effects during the larval stage and may not be captured by metamorphic traits, and (ii) be strongly modulated by egg hatching dates.

Daily temperature variation and extreme high temperatures drive performance and biotic interactions in a warming world.

We review the major patterns on the effects of daily temperature variation (DTV) and extreme high temperatures (EXT) on performance traits and the resulting outcome of biotic interactions in insects. EXT profoundly affects the outcome of all types of biotic interactions: competitive, predator-prey, herbivore-plant, host-pathogen/parasitoid and symbiotic interactions. Studies investigating effects of DTV on biotic interactions are few but also show strong effects on competitive and host-pathogen/parasitoid interactions. EXT typically reduces predation, and is expected to reduce parasitoid success. The effects of EXT and DTV on the outcome of the other interaction types are highly variable, yet can be predicted based on comparisons of the TPCs of the interacting species, and challenges the formulation of general predictions about the change in biotic interactions in a warming world.

Testing the time-scale dependence of delayed interactions: A heat wave during the egg stage shapes how a pesticide interacts with a successive heat wave in the larval stage.

Under global change organisms are exposed to multiple, potentially interacting stressors. Especially interactions between successive stressors are poorly understood and recently suggested to depend on their timing of exposure. We particularly need studies assessing the impact of exposure to relevant stressors at various life stages and how these interact. We investigated the single and combined impacts of a heat wave (mild [25 °C] and extreme [30 °C]) during the egg stage, followed by successive exposure to esfenvalerate (ESF) and a heat wave during the larval stage in damselflies. Each stressor caused mortality. The egg heat wave and larval ESF exposure had delayed effects on survival, growth and lipid peroxidation (MDA). This resulted in deviations from the prediction that stressors separated by a long time interval would not interact: the egg heat wave modulated the interaction between the stressors in the larval stage. Firstly, ESF caused delayed mortality only in larvae that had been exposed to the extreme egg heat wave and this strongly depended upon the larval heat wave treatment. Secondly, ESF only increased MDA in larvae not exposed to the egg heat wave. We found little support for the prediction that when there is limited time between stressors, synergistic interactions should occur. The intermediate ESF concentration only caused delayed mortality when combined with the larval heat wave, and the lowest ESF concentrations only increased oxidative damage when followed by the mild larval heat wave. Survival selection mitigated the interaction patterns between successive stressors that are individually lethal, and therefore should be included in a predictive framework for the time-scale dependence of the outcome of multistressor studies with pollutants. The egg heat wave shaping the interaction pattern between successive pesticide exposure and a larval heat wave highlights the connectivity between the concepts of 'heat-induced pesticide sensitivity' and 'pesticide-induced heat sensitivity'.

Sexual selection reinforces a higher flight endurance in urban damselflies.

Urbanization is among the most important and globally rapidly increasing anthropogenic processes and is known to drive rapid evolution. Habitats in urbanized areas typically consist of small, fragmented and isolated patches, which are expected to select for a better locomotor performance, along with its underlying morphological traits. This, in turn, is expected to cause differentiation in selection regimes, as populations with different frequency distributions for a given trait will span different parts of the species' fitness function. Yet, very few studies considered differentiation in phenotypic traits associated with patterns in habitat fragmentation and isolation along urbanization gradients, and none considered differentiation in sexual selection regimes. We investigated differentiation in flight performance and flight-related traits and sexual selection on these traits across replicated urban and rural populations of the scrambling damselfly Coenagrion puella. To disentangle direct and indirect paths going from phenotypic traits over performance to mating success, we applied a path analysis approach. We report for the first time direct evidence for the expected better locomotor performance in urban compared to rural populations. This matches a scenario of spatial sorting, whereby only the individuals with the best locomotor abilities colonize the isolated urban populations. The covariation patterns and causal relationships among the phenotypic traits, performance and mating success strongly depended on the urbanization level. Notably, we detected sexual selection for a higher flight endurance only in urban populations, indicating that the higher flight performance of urban males was reinforced by sexual selection. Taken together, our results provide a unique proof of the interplay between sexual selection and adaptation to human-altered environments.

The heat is on: Genetic adaptation to urbanization mediated by thermal tolerance and body size.

Worldwide, urbanization leads to tremendous anthropogenic environmental alterations, causing strong selection pressures on populations of animals and plants. Although a key feature of urban areas is their higher temperature ("urban heat islands"), adaptive thermal evolution in organisms inhabiting urban areas has rarely been studied. We tested for evolution of a higher heat tolerance (CTMAX ) in urban populations of the water flea Daphnia magna, a keystone grazer in freshwater ecosystems, by carrying out a common garden experiment at two temperatures (20°C and 24°C) with genotypes of 13 natural populations ordered along a well-defined urbanization gradient. We also assessed body size and haemoglobin concentration to identify underlying physiological drivers of responses in CTMAX . We found a higher CTMAX in animals isolated from urban compared to rural habitats and in animals reared at higher temperatures. We also observed substantial genetic variation in thermal tolerance within populations. Overall, smaller animals were more heat tolerant. While urban animals mature at smaller size, the effect of urbanization on thermal tolerance is only in part caused by reductions in body size. Although urban Daphnia contained higher concentrations of haemoglobin, this did not contribute to their higher CTMAX . Our results provide evidence of adaptive thermal evolution to urbanization in the water flea Daphnia. In addition, our results show both evolutionary potential and adaptive plasticity in rural as well as urban Daphnia populations, facilitating responses to warming. Given the important ecological role of Daphnia in ponds and lakes, these adaptive responses likely impact food web dynamics, top-down control of algae, water quality, and the socio-economic value of urban ponds.