Plastic and evolutionary changes in traits related to biotic interactions are crucial for the local persistence of populations under global warming. Yet, how acute and developmental thermal plasticity evolve and shape predation rates has been poorly studied, especially in the context of latitude-driven thermal evolution. A powerful predictive framework is given by the climatic variability hypothesis (CVH) stating that thermal plasticity and acclimation capacity evolve to be higher in high-latitude populations because these are exposed to higher thermal seasonal variability. We tested the CVH for predation rates and evaluated if the support for the CVH depended on the type of plasticity and acclimation metric. We examined effects of developmental temperature (20 and 24°C) and acute changes in mean and extreme temperatures (20, 24 and 32°C) on the predation rates of high- and low-latitude populations of a predatory aquatic insect, the damselfly Ischnura elegans. We documented opposing and interactive effects between developmental and acute temperatures, which urge caution when using thermal performance curves to forecast the impact of global warming on biotic interactions. Predation rates were higher in low-latitude than high-latitude predators, especially at the warmer developmental and test temperatures, suggesting thermal adaptation to the higher low-latitude temperatures. The latitudinal patterns in acute and developmental plasticities differed, providing mixed support for the CVH. Moreover, there was no latitudinal pattern in post-acclimation thermal sensitivity, indicative of perfect thermal compensation in predators from both latitudes. Strikingly, the acclimation capacity leading to perfect thermal compensation was ~6 times higher in high-latitude than in low-latitude predators. Our study provides new insights into the climatic variability hypothesis (CVH) by documenting that its support is critically dependent on the type of plasticity and acclimation metric used.

Predation rates were collected via 24-hr predation trials with one predator and fixed prey density (90 ind/L) in each trial. Predation rates (consumed) were corrected by body mass (final_mass) and ln-transformed before they were analyzed with linear regression model and customerized contrasts.