Ecologically relevant measures of tolerance to potentially lethal temperatures.

The acute thermal tolerance of ectotherms has been measured in a variety of ways; these include assays where organisms are shifted abruptly to stressful temperatures and assays where organisms experience temperatures that are ramped more slowly to stressful levels. Ramping assays are thought to be more relevant to natural conditions where sudden abrupt shifts are unlikely to occur often, but it has been argued that thermal limits established under ramping conditions are underestimates of true thermal limits because stresses due to starvation and/or desiccation can arise under ramping. These confounding effects might also impact the variance and heritability of thermal tolerance. We argue here that ramping assays are useful in capturing aspects of ecological relevance even though there is potential for confounding effects of other stresses that can also influence thermal limits in nature. Moreover, we show that the levels of desiccation and starvation experienced by ectotherms in ramping assays will often be minor unless the assays involve small animals and last for many hours. Empirical data illustrate that the combined effects of food and humidity on thermal limits under ramping and sudden shifts to stressful conditions are unpredictable; in Drosophila melanogaster the presence of food decreased rather than increased thermal limits, whereas in Ceratitis capitata they had little impact. The literature provides examples where thermal limits are increased under ramping presumably because of the potential for physiological changes leading to acclimation. It is unclear whether heritabilities and population differentiation will necessarily be lower under ramping because of confounding effects. Although it is important to clearly define experimental methods, particularly when undertaking comparative assessments, and to understand potential confounding effects, thermotolerance assays based on ramping remain an important tool for understanding and predicting species responses to environmental change. An important area for further development is to identify the impact of rates of temperature change under field and laboratory conditions.