Evolutionary ecological study of the pace-of-life syndrome in Ischnura damselflies.

To understand and predict patterns in trait variation along environmental gradients and in response to stressors, the integration of traits into suites of correlated traits or ‘syndromes’ has received increasing attention. Because of the presence of such syndromes, the plastic and evolutionary responses of organisms to selection pressures involve coordinated changes in many traits. A promising concept for the integration of traits is the pace-of-life syndrome (POLS) hypothesis, which predicts the integration of life-history, behavioural and physiological traits into a syndrome that varies along a fast-slow continuum. It predicts, for example, that fast-lived animals are characterised by fast growth, early reproduction, a short lifespan, fast metabolism, low investment in body maintenance (such as immune function) and proactive behaviours like high activity and boldness (as opposed to reactive behaviour). Furthermore, such fast-slow variation is expected to exist at different levels of biological organisation: between species, between populations within species and between individuals. In this thesis, I investigated whether there is support for the POLS, how it is affected by different stressors [ultraviolet radiation (UV), contaminants (zinc and the pesticide chlorpyrifos), warming], and whether it can be used to predict organisms’ sensitivity to stress. I did this in a series of common-garden experiments, using coenagrionid damselflies (Ischnura sp. and Coenagrion puella) as model organisms, exploiting the strong variation in life-history strategy linked with voltinism across latitudes and larval habitat types (pond permanence).

I found strong evidence for a fast-slow life-history continuum at all three levels of organisation. Ordering along this continuum at the interspecific and interpopulation level was tightly linked to differences in time constraints imposed by voltinism and larval hydroperiod. Behavioural traits were mostly found to cluster along a proactive-reactive axis, but only at the latitudinal level this was tightly coupled with the fast-slow life-history continuum. The lack of a strong link between life history and behaviour might partly be due to POLS-related patterns in digestive physiology, like conversion and assimilation efficiencies. Patterns in physiology were least in accordance with the predictions under the POLS hypothesis. Among individuals and across latitudes, there was some evidence for the integration of several physiological traits into a syndrome, but this was not consistently associated with a fast-slow continuum. This might be due to the diversity of physiological pathways, that might overlap in function or the existence of trade-offs at smaller trait scales (among two or more physiological traits) instead of trade-offs between pace and investment in body maintenance. In general, the existence of multidimensional trade-offs, where different combinations of traits may lead to a similar pace, might be an explanation for the overall limited integration of traits along a fast-slow axis here observed.

Contaminant exposure (zinc/chlorpyrifos) and warming strongly affected the patterns of covariation among traits at the within-population level. Contaminant exposure generally induced a stronger trait integration among behavioural traits and among physiological traits, indicating stronger trade-offs under energy-constraining conditions. Furthermore the contaminants and temperature treatments (alone and in combination) changed the higher-level associations between life-history, behavioural and physiological traits, leading to coupling, decoupling and inversion of trait associations. Whether these changes affect the fitness of individuals and scale up to impact population dynamics and community structure remain open questions and provide opportunities for future research.

In contrast with the observed plasticity in syndrome structure, within-population trait covariation patterns were remarkably consistent across latitudes, despite strong population-level divergence in trait means associated with local thermal adaptation. This consistency suggests that the trait associations may have acted as evolutionary constraints. If these phenotypic covariation patterns are underpinned by genetic correlations, they might constrain or facilitate future adaptive potential, for example in response to climate change, and this critically depending on the direction of the selection pressure.

Fast-lived low-latitude damselflies were found to be more sensitive to metal contamination than slow-lived high-latitude damselflies. There was also some evidence for a higher contaminant sensitivity in more fast-lived species, but not for a higher sensitivity to UV. These results indicate a potential coupling between POLS and stress sensitivity, yet more studies using a wider array of populations or species are needed to validate this pattern.

In conclusion, I documented an exceptionally large integrated set of life-history, behavioural and physiological traits and covariation patterns within the POLS framework and found the POLS hypothesis predicting trait associations along a unidimensional fast-slow axis of variation to be too simplistic to capture a complex reality. This study further provided a first exploration of how trait covariation patterns might be affected by global warming and pollution. My findings highlight the need for integrated multi-trait approaches, essential to understand and predict how organisms deal with current and future challenges.