Meeting Abstract
Critical temperatures are widely used to quantify the thermal limits of organisms. But measured critical temperatures often vary with methodological details such as starting temperature and ramping rate, leading to spirited discussions about the effects of stress and acclimation during the experiments. We illustrate the statistical reasons why methodological details affect observed values of critical temperature, independent of stress or acclimation. We develop a statistical model that estimates a failure rate function (the relationship between failure rate and current temperature) using maximum likelihood; the best model accounts for 58% of the variation in CTmax in an exemplary dataset for tsetse flies. Our results reject the interpretation of CTmax as a stepwise, threshold temperature. We then extend the model to incorporate potential effects of stress and acclimation on the failure rate function, and apply the model to two insect datasets. The results show how stress accumulation at low ramping rates may shift the failure rate function, decreased observed values of CTmax (tsetse flies); and how acclimation (heat-hardening) may reduce the slope of the failure rate function, increasing observed values of CTmax (hornworm larvae). The model provides a new approach to analyzing and interpreting critical temperatures.
