Temperature Manipulation Induces Ecdysis in Lab-Held Rattlesnakes
Ecdysis is a defining feature of all squamate reptiles, occurring at least annually in snakes. Shedding occurs at irregular and unpredictable intervals, inhibiting studies of the importance of ecdysis to life history. To date, limited information on the endocrine mechanisms controlling ecdysis has been used to induce sheds via endocrine gland removal. However, whole-gland removal is often lethal, limiting the scope of any conclusions drawn from these studies. We used data from the literature and reports of the natural timing of shed events in rattlesnakes to design a novel, non-invasive method for inducing ecdysis in the laboratory. Temperature manipulations following simulated brumation periods were used to induce ecdysis. Temperature trials were successful in inducing ecdysis in 63% of treated animals (28 of 44). Efficacy of the induction of ecdysis via temperature manipulation was related to body condition and time since previous shed cycle (with high body condition animals that had not shed in the three months prior being most likely to shed). The timing of ecdysis (Immediate or Delayed) relative to temperature manipulation was related to days since last shed but not body condition, with those who had not shed > 5 months prior to manipulations being most likely to shed immediately after warming from brumation temperatures (10°C). Temperature manipulation shows potential for use in the lab-based induction and study of ecdysis, particularly in high body condition animals that have not shed in several months. A reliable method to induce ecdysis in the laboratory provides a novel opportunity to investigate the physiological and ecological implications of an under-studied but ubiquitous function in squamates.
Shed group (No Shed, Delayed, or Immediate) by time (days elapsed between last shed event and end of temperature increase). Point shape and color denote species (blue triangles = C. viridis, red circles = C. horridus), and shading intensity corresponds to calculated body condition index (BCI), with darker shaded shapes indicating animals in “good” body condition compared to others. There was a difference between groups with sheds being delayed or prevented in animals that had shed within ∼150 days of the trial end.
Heat map of the likelihood of observing a shed in a given three-week period of the year (top distribution; Control), compared to the proportions of animals that shed after temperature trials (Trial A; bottom distribution, Trial B; top distribution). Darker purple indicates higher shed likelihood/proportion; light orange squares indicate periods where sheds were rarely observed. Note that sheds increase immediately following Trial A and decrease with time, where Trial B animals tended to shed with a delay following treatment. Additionally, the imperfect division of the year into 21-day (three-week) blocks results in a short two-week period in late December.
Binary logistic regression of all snakes (n = 44) showing the probability (left y-axes) that individuals shed (Shed vs. No Shed) in response to temperature treatment. Plots show likelihood of shedding as a function of time since last shed cycle (left panel) or estimated body condition index at the start of temperature manipulation (BCI; right panel). Histograms show number of observations (right y-axes, “frequency”) in each bin (bins of 50 days for time since shed; bin increments of 0.05 BCI for body condition) for animals that did not shed in response to manipulations (bottom distributions, left and right) and snakes that did shed (top distributions, left and right). Both time since last shed and body condition were significant predictors of likelihood of shed (P values < 0.05), indicating that animals in good condition, with long delays since last shed, or individuals fitting both conditions were most likely to respond favorably to temperature manipulation.
Binary logistic regression of snakes that shed in response to temperature manipulation (n = 28) showing the likelihood of that shed response occurring immediately after (bottom distributions) or with a temporal delay (top distributions). There was a significant effect of time since last shed (left panel) but not body condition (right panel) on the timing of shed events. Animals that did not shed were not considered in this analysis but were included in an alternate binary regression (Fig. 3). Time since last shed but not body condition was a significant predictor of likelihood of timing of shed (P < 0.005), indicating that animals that have not shed in over ∼5 months were most likely to shed immediately in response to temperature manipulations.
Contributor Notes
Associate Editor: C. Bevier