Maintenance of Behavioral Variation under Predation Risk: Effects on Personality, Plasticity, and Predictability.

AbstractClassic evolutionary theory predicts that predation will shift trait means and erode variance within prey species; however, several studies indicate higher behavioral trait variance and trait integration in high-predation populations. These results come predominately from field-sampled animals comparing low- and high-predation sites and thus cannot isolate the role of predation from other ecological factors, including density effects arising from higher predation. Here, we study the role of predation on behavioral trait (co)variation in experimental populations of guppies (Poecilia reticulata) living with and without a benthic ambush predator (Jaguar cichlid) to better evaluate the role of predation and where density was equalized among replicates twice per year. At 2.5 years after introduction of the predators (∼10 overlapping generations), 40 males were sampled from each of the six replicate populations and extensively assayed for activity rates, water column use, and latency to feed following disturbance. Individual variation was pronounced in both treatments, with substantial individual variation in means, temporal plasticity, and predictability (inverse residual variance). Predators had little effect on mean behavior, although there was some evidence for greater use of the upper water column in predator-exposed fish. There was greater variance among individuals in water column use in predator-exposed fish, and they habituated more quickly over time; individuals higher in the water column fed slower and had a reduced positive correlation with activity, although again this effect was time specific. Predators also affected the integration of personality and plasticity-among-individual variances in water column use increased, and those in activity decreased, through time-which was absent in controls. Our results contrast with the extensive guppy literature showing rapid evolution in trait means, demonstrating either increases or maintenance of behavioral variance under predation.

Behavioural ecology meets oncology: quantifying the recovery of animal behaviour to a transient exposure to a cancer risk factor.

Wildlife is increasingly exposed to sublethal transient cancer risk factors, including mutagenic substances, which activates their anti-cancer defences, promotes tumourigenesis, and may negatively impact populations. Little is known about how exposure to cancer risk factors impacts the behaviour of wildlife. Here, we investigated the effects of a sublethal, short-term exposure to a carcinogen at environmentally relevant concentrations on the activity patterns of wild Girardia tigrina planaria during a two-phase experiment, consisting of a 7-day exposure to cadmium period followed by a 7-day recovery period. To comprehensively explore the effects of the exposure on activity patterns, we employed the double hierarchical generalized linear model framework which explicitly models residual intraindividual variability in addition to the mean and variance of the population. We found that exposed planaria were less active compared to unexposed individuals and were able to recover to pre-exposure activity levels albeit with a reduced variance in activity at the start of the recovery phase. Planaria showing high activity levels were less predictable with larger daily activity variations and higher residual variance. Thus, the shift in behavioural variability induced by an exposure to a cancer risk factor can be quantified using advanced tools from the field of behavioural ecology. This is required to understand how tumourous processes affect the ecology of species.

Testing personality-pace-of-life associations via artificial selection: insights from selected lines of rainbow trout on the context-dependence of correlations.

More than a decade of study since the personality pace-of-life syndrome (POLS) hypotheses were first proposed, there is little support for it within species. Lack of experimental control, insufficient sampling in the face of highly labile behavioural and metabolic traits, and context dependency of trait correlations are suggested as reasons. Here, I argue that artificial selection and/or use of existing selected lines represents a powerful but under-used approach to furthering our understanding of the POLS. To illustrate this potential, I conducted a focussed review of studies that compared the behaviour, metabolism, growth and survival of an artificially selected fast-growing rainbow trout relative to wild unselected strains, under varying food and risk conditions in the laboratory and field. Resting metabolic rate, food intake, and behaviours that enhance feeding but increase energy expenditure (activity, aggression, boldness), were all higher in the fast strain in paired contrasts, under all food and risk conditions, both in the laboratory and the field. Fast-strain fish grew faster in almost every food and risk situation except where food was highly limited (or absent), had higher survival under low or zero predation risk, but had lower survival under high risk. Several other traits rarely considered in POLS studies were also higher in the fast strain, including maximum swimming speed, and hormones (growth hormone (GH), thyroid hormone (T3) and insulin-like growth factor (IGF-1)). I conclude: (i) assumptions and predictions of the POLS hypothesis are well supported, and (ii) context-dependency was largely absent, but when present revealed trade-offs between food acquisition and predation risk. This focused review highlights the potential of artificial selection in testing POLS ideas, and will hopefully motivate further studies using other animals.

Meta-analytic insights into factors influencing the repeatability of hormone levels in agricultural, ecological, and medical fields.

Interest in individual variation in hormone concentrations is rapidly increasing, particularly with regard to the evolutionary and practical implications. A key aspect of studying individual variation in any labile trait is estimating the degree of within- versus among-individual variation, but at present, we do not have a broad consensus on the extent to which hormone levels are repeatable and what factors might influence repeatability. To address this knowledge gap, we conducted a comprehensive meta-analysis of hormone levels that included 1,132 estimates of repeatability from 368 studies across three fields of study: agriculture, ecology, and medicine. We assessed the influence of sex, age class, sample type, hormone family, type of hormone measure, assay type, number of subjects, number of samples per subject, and sampling interval on repeatability estimates. Overall mean repeatability was 0.58, but estimates differed substantially among study disciplines, being lowest in ecology (0.34), moderate in agriculture (0.52), and relatively high in medicine (0.68). In addition, repeatability decreased slightly as sampling interval increased, and also tended to be higher for peak hormone levels than baseline levels. Overall, hormone levels are moderately repeatable, suggesting that they can potentially be useful indicators of individual variation. However, estimates of repeatability are quite variable among fields, so caution should be used when relying on single samples to assess individual variation.

The influence of environmental gradients on individual behaviour: Individual plasticity is consistent across risk and temperature gradients.

The expression of individual behaviour as a function of environmental variation (behavioural plasticity) is recognized as a means for animals to modify their phenotypes in response to changing conditions. Plasticity has been studied extensively in recent years, leading to an accumulation of evidence for behavioural plasticity within natural populations. Despite the recent attention given to studying individual variation in behavioural plasticity, there is still a lack of consensus regarding its causes and constraints. One pressing question related to this is whether individual plasticity carries over across temporal and environmental gradients. That is, are some individuals more plastic (responsive) than others in general? Here, we examined the influence of temporal and environmental gradients on individual behavioural responses in a marine gastropod, Littoraria irrorata. We measured individual boldness repeatedly over time and in response to tidal cycle (high vs. low, an index of risk) and daily temperature fluctuations (known to affect metabolism), in a controlled field experiment. On average, boldness increased from high to low tide and with increasing temperature but decreased marginally over time. Individuals also differed in their responses to variation in tide and temperature, but not over time. Those which were relatively bold at high tide (when predation risk is greater) were similarly bold at low tide, whereas shy individuals became much more "bold" at low tide. Most notably, individuals that were more responsive to tide (and thus risk) were also more responsive to temperature changes, indicating that plasticity was correlated across contexts (r = 0.57) and that bolder individuals were least plastic overall. This study provides a rare and possibly first example of consistency of individual behavioural plasticity across contexts, suggesting underlying physiology as a common mechanism, and raises the possibility of correlational selection on plasticity.

Metabolic Scope as a Proximate Constraint on Individual Behavioral Variation: Effects on Personality, Plasticity, and Predictability.

Behavioral ecologists have hypothesized that among-individual differences in resting metabolic rate (RMR) may predict consistent individual differences in mean values for costly behaviors or for behaviors that affect energy intake rate. This hypothesis has empirical support and presently attracts considerable attention, but, notably, it does not provide predictions for individual differences in (a) behavioral plasticity or (b) unexplained variation (residual variation from mean individual behavior, here termed predictability). We outline how consideration of aerobic maximum metabolic rate (MMR) and particularly aerobic scope (= MMR - RMR) can be used to simultaneously make predictions about mean and among- and within-individual variation in behavior. We predict that while RMR should be proportional to an individual's mean level of sustained behavioral activity (one aspect of its personality), individuals with greater aerobic scope will also have greater scope to express behavioral plasticity and/or greater unpredictability in behavior (=greater residual variation). As a first step toward testing these predictions, we analyze existing activity data from selectively bred lines of mice that differ in both daily activity and aerobic scope. We find that replicate high-scope mice are more active on average and show greater among-individual variation in activity, greater among-individual variation in plasticity, and greater unpredictability. These data provide some tentative first support for our hypothesis, suggesting that further research on this topic would be valuable.

Is behavioural plasticity consistent across different environmental gradients and through time?

Despite accumulating evidence for individual variation in behavioural plasticity, there is currently little understanding of the causes and consequences of this variation. An outstanding question is whether individual reaction norm (RN) slopes are consistent across different environmental variables-that is, whether an individual that is highly responsive to one environmental variable will be equally responsive to a second variable. Another important and related question is whether RNs are themselves consistently expressed through time or whether they are simply state dependent. Here, we quantified individual activity rates of zebrafish in response to independent manipulations of temperature and food availability that were repeated in discrete 'bursts' of sampling through time. Individuals that were thermally responsive were not more responsive to food deprivation, but they did exhibit greater unexplained variation. Individual RN slopes were consistent (repeatable) over time for both temperature (Rslope = 0.92) and food deprivation responses (Rslope = 0.4), as were mean activity rates in the standard environment (Rintercept = 0.83). Despite the high potential lability of behaviour, we have demonstrated consistency of behavioural RN components and identified potential energetic constraints leading to high consistency of thermal RNs and low consistency of food deprivation RNs.

Cancer and life-history traits: lessons from host-parasite interactions.

Despite important differences between infectious diseases and cancers, tumour development (neoplasia) can nonetheless be closely compared to infectious disease because of the similarity of their effects on the body. On this basis, we predict that many of the life-history (LH) responses observed in the context of host-parasite interactions should also be relevant in the context of cancer. Parasites are thought to affect LH traits of their hosts because of strong selective pressures like direct and indirect mortality effects favouring, for example, early maturation and reproduction. Cancer can similarly also affect LH traits by imposing direct costs and/or indirectly by triggering plastic adjustments and evolutionary responses. Here, we discuss how and why a LH focus is a potentially productive but under-exploited research direction for cancer research, by focusing our attention on similarities between infectious disease and cancer with respect to their effects on LH traits and their evolution. We raise the possibility that LH adjustments can occur in response to cancer via maternal/paternal effects and that these changes can be heritable to (adaptively) modify the LH traits of their offspring. We conclude that LH adjustments can potentially influence the transgenerational persistence of inherited oncogenic mutations in populations.

Fishing directly selects on growth rate via behaviour: implications of growth-selection that is independent of size.

Size-selective harvest of fish and crustacean populations has reduced stock numbers, and led to reduced growth rates and earlier maturation. In contrast to the focus on size-selective effects of harvest, here, we test the hypothesis that fishing may select on life-history traits (here, growth rate) via behaviour, even in the absence of size selection. If true, then traditional size-limits used to protect segments of a population cannot fully protect fast growers, because at any given size, fast-growers will be more vulnerable owing to bolder behaviour. We repeatedly measured individual behaviour and growth of 86 crayfish and found that fast-growing individuals were consistently bold and voracious over time, and were subsequently more likely to be harvested in single- and group-trapping trials. In addition, there was some indication that sex had independent effects on behaviour and trappability, whereby females tended to be less active, shyer, slower-growing and less likely to be harvested, but not all these effects were significant. This study represents, to our knowledge, the first across-individual support for this hypothesis, and suggests that behaviour is an important mechanism for fishing selectivity that could potentially lead to evolution of reduced intrinsic growth rates.

Individual and sex-specific differences in intrinsic growth rate covary with consistent individual differences in behaviour.

The evolutionary causes of consistent individual differences in behaviour are currently a source of debate. A recent hypothesis suggests that consistent individual differences in life-history productivity (growth and/or fecundity) may covary with behavioural traits that contribute to growth-mortality trade-offs, such as risk-proneness (boldness) and foraging activity (voraciousness). It remains unclear, however, to what extent individual behavioural and life-history profiles are set early in life, or are a more flexible result of specific environmental or developmental contexts that allow bold and active individuals to acquire more resources. Longitudinal studies of individually housed animals under controlled conditions can shed light on this question. Since growth and behaviour can both vary within individuals (they are labile), studying between-individual correlations in behaviour and growth rate requires repeated scoring for both variables over an extended period of time. However, such a study has not yet been done. Here, we repeatedly measured individual mass seven times each, boldness 40 times each and voracity eight times each during the first 4 months of life on 90 individually housed crayfish (Cherax destructor). Animals were fed ad libitum, generating a context where individuals can express their intrinsic growth rate (i.e. growth capacity), but in which bold and voracious behaviour is not necessary for high resource acquisition (crayfish can and do hoard food back to their burrow). We show that individuals that were consistently bold over time during the day were also bolder at night, were more voracious and maintained higher growth rates over time than shy individuals. Independent of individual differences, we also observed that males were faster-growing, bolder and more voracious than females. Our findings imply that associations between bold behaviour and fast growth can occur in unlimited food contexts where there is no necessary link between bold behaviour and resource acquisition - offering support for the 'personality-productivity' hypothesis. We suggest future research should study links between consistent individual differences in behaviour and life history under a wider range of contexts, in order to shed light on the role of biotic and abiotic conditions in the strength, direction and stability of their covariance.

Individual variation in thermal performance curves: swimming burst speed and jumping endurance in wild-caught tropical clawed frogs.

The importance of studying individual variation in locomotor performance has long been recognized as it may determine the ability of an organism to escape from predators, catch prey or disperse. In ectotherms, locomotor performance is highly influenced by ambient temperature (Ta), yet several studies have showed that individual differences are usually retained across a Ta gradient. Less is known, however, about individual differences in thermal sensitivity of performance, despite the fact that it could represent adaptive sources of phenotypic variation and/or additional substrate for selection to act upon. We quantified swimming and jumping performance in 18 wild-caught tropical clawed frogs (Xenopus tropicalis) across a Ta gradient. Maximum swimming velocity and acceleration were not repeatable and individuals did not differ in how their swimming performance varied across Ta. By contrast, time and distance jumped until exhaustion were repeatable across the Ta gradient, indicating that individuals that perform best at a given Ta also perform best at another Ta. Moreover, thermal sensitivity of jumping endurance significantly differed among individuals, with individuals of high performance at low Ta displaying the highest sensitivity to Ta. Individual differences in terrestrial performance increased with decreasing Ta, which is opposite to results obtained in lizards at the inter-specific and among-individual levels. To verify the generality of these patterns, we need more studies on individual variation in thermal reaction norms for locomotor performance in lizards and frogs.

The effect of mitochondrial recombination on fertilization success in blue mussels.

The presence of doubly uniparental inheritance (DUI) in bivalves represents a unique mode of mitochondrial transmission, whereby paternal (male-transmitted M-type) and maternal (female-transmitted F-type) haplotypes are transmitted to offspring separately. Male embryos retain both haplotypes, but the M-type is selectively removed from females. Due to the presence of heteroplasmy in males, mtDNA can recombine resulting in a 'masculinized' haplotype referred to as Mf-type. While mtDNA recombination is usually rare, it has been recorded in multiple mussel species across the Northern Hemisphere. Given that mitochondria are the powerhouse of the cell, different mtDNA haplotypes may have different selective advantages under diverse environmental conditions. This may be particularly important for sperm fitness and fertilization success. In this study we aimed to i) determine the presence, prevalence of the Mf-type in Australian blue mussels (Mytilus sp.) and ii) investigate the effect of Mf-mtDNA on sperm performance (a fitness correlate). We found a high prevalence of recombined mtDNA (≈35 %) located within the control region of the mitochondrial genome, which occurred only in specimens that contained Southern Hemisphere mtDNA. The presence of two female mitotypes were identified in the studied mussels, one likely originating from the Northern Hemisphere, and the other either representing the endemic M. planulatus species or introduced genotypes from the Southern Hemisphere. Despite having recombination events present in a third of the studied population, analysis of sperm performance indicated no difference in fertilization success related to mitotype.

Predictability as a personality trait: consistent differences in intraindividual behavioral variation.

Although animal behavior is generally repeatable, most behavioral variation apparently occurs within rather than across individuals. With the exception of very recent interest in individual behavioral plasticity (consistent differences in responsiveness), this within-individual variation has been largely ignored despite its importance in the study of proximate and ultimate questions about behavior. Here, we repeatedly scored the undisturbed activity of 30 adult male mosquitofish across multiple observation bouts spanning 132 days ([Formula: see text] observations per fish). We found that the behavior of some individuals was consistently more predictable in a given context than others. Repeatability for this "intraindividual variation" (IIV; [Formula: see text]) was evident after accounting for individual differences in activity trends across days, and activity responses due to fine-scale temperature variation (i.e., individual plasticity in response to both variables). To our knowledge, this is the first evidence that predictability of behavior is a repeatable characteristic of individual animals. We suggest that IIV represents an important axis of consistent behavioral variation that has previously not been formally considered. Finally, individual differences in predictability may similarly exist for labile morphological and physiological traits but have seemingly not been studied.

Are most samples of animals systematically biased? Consistent individual trait differences bias samples despite random sampling.

Sampling animals from the wild for study is something nearly every biologist has done, but despite our best efforts to obtain random samples of animals, 'hidden' trait biases may still exist. For example, consistent behavioral traits can affect trappability/catchability, independent of obvious factors such as size and gender, and these traits are often correlated with other repeatable physiological and/or life history traits. If so, systematic sampling bias may exist for any of these traits. The extent to which this is a problem, of course, depends on the magnitude of bias, which is presently unknown because the underlying trait distributions in populations are usually unknown, or unknowable. Indeed, our present knowledge about sampling bias comes from samples (not complete population censuses), which can possess bias to begin with. I had the unique opportunity to create naturalized populations of fish by seeding each of four small fishless lakes with equal densities of slow-, intermediate-, and fast-growing fish. Using sampling methods that are not size-selective, I observed that fast-growing fish were up to two-times more likely to be sampled than slower-growing fish. This indicates substantial and systematic bias with respect to an important life history trait (growth rate). If correlations between behavioral, physiological and life-history traits are as widespread as the literature suggests, then many animal samples may be systematically biased with respect to these traits (e.g., when collecting animals for laboratory use), and affect our inferences about population structure and abundance. I conclude with a discussion on ways to minimize sampling bias for particular physiological/behavioral/life-history types within animal populations.

Time-specific convergence and divergence in individual differences in behavior: Theory, protocols and analyzes.

Over the years, theoreticians and empiricists working in a wide range of disciplines, including physiology, ethology, psychology, and behavioral ecology, have suggested a variety of reasons why individual differences in behavior might change over time, such that different individuals become more similar (convergence) or less similar (divergence) to one another. Virtually none of these investigators have suggested that convergence or divergence will continue forever, instead proposing that these patterns will be restricted to particular periods over the course of a longer study. However, to date, few empiricists have documented time-specific convergence or divergence, in part because the experimental designs and statistical methods suitable for describing these patterns are not widely known. Here, we begin by reviewing an array of influential hypotheses that predict convergence or divergence in individual differences over timescales ranging from minutes to years, and that suggest how and why such patterns are likely to change over time (e.g., divergence followed by maintenance). Then, we describe experimental designs and statistical methods that can be used to determine if (and when) individual differences converged, diverged, or were maintained at the same level at specific periods during a longitudinal study. Finally, we describe why the concepts described herein help explain the discrepancy between what theoreticians and empiricists mean when they describe the "emergence" of individual differences or personality, how they might be used to study situations in which convergence and divergence patterns alternate over time, and how they might be used to study time-specific changes in other attributes of behavior, including individual differences in intraindividual variability (predictability), or genotypic differences in behavior.

The impact of food availability on tumorigenesis is evolutionarily conserved.

The inability to control cell proliferation results in the formation of tumors in many multicellular lineages. Nonetheless, little is known about the extent of conservation of the biological traits and ecological factors that promote or inhibit tumorigenesis across the metazoan tree. Particularly, changes in food availability have been linked to increased cancer incidence in humans, as an outcome of evolutionary mismatch. Here, we apply evolutionary oncology principles to test whether food availability, regardless of the multicellular lineage considered, has an impact on tumorigenesis. We used two phylogenetically unrelated model systems, the cnidarian Hydra oligactis and the fish Danio rerio, to investigate the impact of resource availability on tumor occurrence and progression. Individuals from healthy and tumor-prone lines were placed on four diets that differed in feeding frequency and quantity. For both models, frequent overfeeding favored tumor emergence, while lean diets appeared more protective. In terms of tumor progression, high food availability promoted it, whereas low resources controlled it, but without having a curative effect. We discuss our results in light of current ideas about the possible conservation of basic processes governing cancer in metazoans (including ancestral life history trade-offs at the cell level) and in the framework of evolutionary medicine.

A novel perspective suggesting high sustained energy expenditure may be net protective against cancer.

Energy expenditure (EE) is generally viewed as tumorigenic, due to production of reactive oxygen species (ROS) that can damage cells and DNA. On this basis, individuals within a species that sustain high EE should be more likely to develop cancer. Here, we argue the opposite, that high EE may be net protective effect against cancer, despite high ROS production. This is possible because individuals that sustain high EE have a greater energetic capacity (=greater energy acquisition, expenditure and ability to up-regulate output), and can therefore allocate energy to multiple cancer-fighting mechanisms with minimal energetic trade-offs. Our review finds that individuals sustaining high EE have greater antioxidant production, lower oxidative stress, greater immune function and lower cancer incidence. Our hypothesis and literature review suggest that EE may indeed be net protective against cancer, and that individual variation in energetic capacity may be a key mechanism to understand the highly individual nature of cancer risk in contemporary human populations and laboratory animals. Lay summary The process of expending energy generates reactive oxygen species that can lead to oxidative stress, cell and DNA damage, and the accumulation of this damage is thought to be a major contributor to many ageing related diseases that include cancer. Here, we challenge this view, proposing how and why high energy expenditure (EE) may actually be net protective against cancer, and provide literature support for our hypothesis. We find individuals with high sustained EE have greater energetic capacity and thus can invest more in repair to counter oxidative stress, and more in immune function, both of which reduce cancer risk. Our hypothesis provides a novel mechanism to understand the highly individual nature of cancer, why taller individuals are more at risk, why physically active individuals have lower cancer risk, and why regular exercise can reduce cancer risk.

Rapid depletion of genotypes with fast growth and bold personality traits from harvested fish populations.

The possibility for fishery-induced evolution of life history traits is an important but unresolved issue for exploited fish populations. Because fisheries tend to select and remove the largest individuals, there is the evolutionary potential for lasting effects on fish production and productivity. Size selection represents an indirect mechanism of selection against rapid growth rate, because individual fish may be large because of rapid growth or because of slow growth but old age. The possibility for direct selection on growth rate, whereby fast-growing genotypes are more vulnerable to fishing irrespective of their size, is unexplored. In this scenario, faster-growing genotypes may be more vulnerable to fishing because of greater appetite and correspondingly greater feeding-related activity rates and boldness that could increase encounter with fishing gear and vulnerability to it. In a realistic whole-lake experiment, we show that fast-growing fish genotypes are harvested at three times the rate of the slow-growing genotypes within two replicate lake populations. Overall, 50% of fast-growing individuals were harvested compared with 30% of slow-growing individuals, independent of body size. Greater harvest of fast-growing genotypes was attributable to their greater behavioral vulnerability, being more active and bold. Given that growth is heritable in fishes, we speculate that evolution of slower growth rates attributable to behavioral vulnerability may be widespread in harvested fish populations. Our results indicate that commonly used minimum size-limits will not prevent overexploitation of fast-growing genotypes and individuals because of size-independent growth-rate selection by fishing.

Understanding the unexplained: The magnitude and correlates of individual differences in residual variance.

Behavioral and physiological ecologists have long been interested in explaining the causes and consequences of trait variation, with a focus on individual differences in mean values. However, the majority of phenotypic variation typically occurs within individuals, rather than among individuals (as indicated by average repeatability being less than 0.5). Recent studies have further shown that individuals can also differ in the magnitude of variation that is unexplained by individual variation or environmental factors (i.e., residual variation). The significance of residual variation, or why individuals differ, is largely unexplained, but is important from evolutionary, methodological, and statistical perspectives. Here, we broadly reviewed literature on individual variation in behavior and physiology, and located 39 datasets with sufficient repeated measures to evaluate individual differences in residual variance. We then analyzed these datasets using methods that permit direct comparisons of parameters across studies. This revealed substantial and widespread individual differences in residual variance. The magnitude of individual variation appeared larger in behavioral traits than in physiological traits, and heterogeneity was greater in more controlled situations. We discuss potential ecological and evolutionary implications of individual differences in residual variance and suggest productive future research directions.

Small within-day increases in temperature affects boldness and alters personality in coral reef fish.

Consistent individual differences in behaviour, termed personality, are common in animal populations and can constrain their responses to ecological and environmental variation, such as temperature. Here, we show for the first time that normal within-daytime fluctuations in temperature of less than 3 degrees C have large effects on personality for two species of juvenile coral reef fish in both observational and manipulative experiments. On average, individual scores on three personality traits (PTs), activity, boldness and aggressiveness, increased from 2.5- to sixfold as a function of temperature. However, whereas most individuals became more active, aggressive and bold across temperature contexts (were plastic), others did not; this changed the individual rank order across temperatures and thus altered personality. In addition, correlations between PTs were consistent across temperature contexts, e.g. fish that were active at a given temperature also tended to be both bold and aggressive. These results (i) highlight the importance of very carefully controlling for temperature when studying behavioural variation among and within individuals and (ii) suggest that individual differences in energy metabolism may contribute to animal personality, given that temperature has large direct effects on metabolic rates in ectotherms.