Summary Over the coming decades, our planet will experience a dramatic increase in average temperatures and an increase in the variance around those temperatures leading to more frequent and harsher heat waves. These changes will impact most species and impose strong selection on physiological traits. Rapid acclimation is the most direct way for organisms to respond to such extreme events, but we currently have little understanding of how the capacity to mount such plastic responses evolves. Accordingly, there is some urgency to determine how the physiological response to high temperatures varies within species, and how this variation is driven by the environment. Here, we investigate heat‐hardening capacity – a rapid physiological response that confers a survival advantage under extreme thermal stress – across 13 populations of a rain forest lizard, Lampropholis coggeri, from the tropics of north‐eastern Australia. Our results reveal that heat hardening is constrained in these lizards by a hard upper thermal limit for locomotor function (approximately 43 °C). Further, hardening response shows strong geographic variation associated with thermal environment: lizards from more predictable and more seasonal thermal environments exhibited greater hardening compared with those from more stochastic and less seasonal habitats. This finding – that predictability in thermal variation influences hardening capacity – aligns closely with theoretical expectations. Our results suggest that tropical species may harbour adaptive variation in physiological plasticity that they can draw from in response to climate change, and this variation is spatially structured in locally adapted populations. Our results also suggest that, by using climatic data, we can predict which populations contain particular adaptive variants; information critical to assisted gene flow strategies.

Aim The exceptionally rich biodiversity found in tropical rainforest is under threat from anthropogenic climate change. We recognize the threat, yet we have little knowledge of the capacity of tropical species to adjust their climate sensitivity in response to it. One indicator of a species’ capacity to adjust to different climates is the amount of intraspecific variation observed in its climate‐relevant traits; if a climate‐relevant trait varies, and this variation is correlated with local climates, it suggests the species can adjust the trait to different conditions through either phenotypic plasticity or evolutionary adaptation. Here, we test for intraspecific variation in climate‐relevant traits in a rainforest specialist to shed light on the capacity of such species to adjust to different climates. Location The Wet Tropics Bioregion, Australia. Methods We studied 12 populations of a lizard that is a tropical rainforest specialist, the rainforest sunskink (Lampropholis coggeri), testing for intraspecific variation in four traits that are potentially important in determining a species’ climate sensitivity. The measured traits were as follows: critical thermal minimum, critical thermal maximum, thermal optimum for sprinting, and desiccation rate. Results We found substantial variation both through time and across space in the measured traits, suggesting both strong plasticity and substantial geographic variation. Moreover, trait variation was correlated with local climate variables, suggesting variation reflects adjustment to local conditions. Main conclusions If physiological lability similar to that observed in rainforest sunskinks occurs in tropical rainforest species more generally, these taxa may not be as climatically specialized, and so not as vulnerable to climate change, as previously thought.

The two foremost hypotheses on the evolutionary constraints on an organism's thermal sensitivity -the hotter-is-better expectation, and the specialistgeneralist trade-off -have received mixed support from empirical studies testing for their existence. Could these conflicting results reflect confusion regarding the organizational level (i.e. species > population > individual) at which these constraints should manifest? We propose that these evolutionary constraints should manifest at different organizational levels because of differences in their underlying causes and requirements. The hotter-is-better expectation should only manifest across separate evolutionary units (e.g. species, populations), and not within populations. The specialist-generalist trade-off, by contrast, should manifest within as well as between separate evolutionary units. We measured the thermal sensitivity of sprint performance for 440 rainforest sun skinks (Lampropholis coggeri) representing 10 populations, and used the resulting performance curves to test for evidence for the hypothesized constraints at two organizational levels: (i) across populations and (ii) within populations. As predicted, the hotter-is-better expectation was evident only at the across-population level, whereas the specialist-generalist trade-off was evident within, as well as across, populations. Our results suggest that, depending on the processes that drive them, evolutionary constraints can manifest at different organizational levels. Consideration of these underlying processes, and the organizational level at which a constraint should manifest, may help resolve conflicting empirical results.