The effect of thermal gradient design on the evaluation of thermoregulation in snakes

Hertz et al. (1993) designed what is now the most widely used protocol to analyse the thermal strategies and efficiency of small squamates. Preferred temperature range (Tp) is one of the most important variables required for determining the thermal efficiency index, and is calculated by monitoring the body temperature of the individuals in an enclosure containing a thermal gradient. Although thermoregulation studies of lizards have traditionally employed thermal gradients under laboratory conditions, this approach is not suitable for snakes given that such thermal gradients do not accurately represent their natural thermal environment and thus may result in snakes selecting suboptimal temperatures. Here, we compare the results of this thermal efficiency protocol using a laboratory thermal gradient (LG) and a semi-captivity thermal gradient (SCG) in the rattlesnake Crotalus polystictus. We found traces of seasonal variation in the SCG Tp, but this could not be assessed in the LG. Tp from the LG was much higher (29 – 34.3 °C) than from the SCG (22.5-30.9 °C). Values for the accuracy of thermoregulation (db) and thermal quality of the environment (de) indices from the LG were consistently higher than from the SCG. However, the efficiency of thermoregulation (E) was higher when calculated from the SCG. Tp estimates were wider than most that have been obtained from other snake species, suggesting that C. polystictus is eurythermic. The Blowin Demers and Weatherhead index was nearly identical in both gradients. Results from the LG indicated that C. polystictus is an inaccurate and inefficient thermoregulator, due to the higher temperatures chosen in this environment. In contrast, results from the SCG suggested that it is a highly accurate and active thermoregulator. We suggest that the LG could represent a stressful environment for snakes, and, as a consequence, they might select higher temperatures to increase anti-predatory performance at the expense of less efficient thermoregulation. Generally, a thermal gradient that more accurately replicates the natural habitat of snake species should reduce stress and result in more robust estimates of thermoregulatory variables.