1. In thermally extreme environments, it is challenging for organisms to maximize performance due to risks associated with stochastic variation in temperature and, subsequently, over evolutionary time minimizing the exposure to risk can serve as one of the mechanisms that result in organisms preferring suboptimal temperatures.
2. We tested this hypothesis in a slow-moving intertidal snail on tropical rocky shores, where temperature variability increases with time from 30 min to 20 hr when recorded at 30 min intervals (due to short-term environmental autocorrelation where temperatures closer in time are more similar as compared to temperatures over a long period of time). Failure to accommodate temporal variation in thermal stress by selecting cool habitats can result in mortality.
3. Thermal performance curves for different traits (heart rate and locomotion) were measured and compared to the snail's thermal preferences in both the field and laboratory. Predicted performances of the snails were simulated based on thermal performance curves for different traits over multiple time-scales and simulated carryover effects.
4. A strong mismatch was found between physiological and behavioural thermal maxima of the snails (physiological thermal maximum being higher by ~7°C), but the snails avoided these maxima and sought temperatures 7–14°C cooler. Such a risk-averse strategy can be explained by their predicted performances where the snails should make decisions about preferred temperatures based on time periods ≥5 hr to avoid underestimating the temporal variation in body temperature.
5. In extreme and stochastic environments, where the temporal variation in environmental conditions can lead to substantial divergence between instantaneous and time-averaged thermal performances, ‘cooler is better’ and ‘suboptimal’ body temperatures are preferred as they provide sufficient buffer to reduce mortality risk from heat stress.
Bibliographical noteThe authors thank Martin Cheng and Alex Li for assistance in the field and laboratory, and Sarah Lau for discussion in improving the manuscript. They also thank four reviewers, including D. Bolnick and L. Buckley for earlier comments on the manuscript. S.C. was supported by General Research Fund 17138916 awarded to G.A.W from the Research Grants Council of Hong Kong. This research was supported by the ProTherm Division Grant from the Division of Ecology & Biodiversity, The University of Hong Kong.
© 2022 The Authors. Journal of Animal Ecology © 2022 British Ecological Society.
- Jensen's inequality
- preferred temperature
- thermal performance