Effects of aging on timing of hibernation and reproduction

Abstract

Small hibernators are long-lived for their size because seasonal dormancy greatly reduces predation risk. Thus, within a year, hibernators switch between states of contrasting mortality risk (active season versus hibernation), making them interesting species for testing the predictions of life-history theory. Accordingly, we hypothesized that, with advancing age and hence diminishing reproductive potential, hibernators should increasingly accept the higher predation risk associated with activity to increase the likelihood of current reproductive success. For edible dormice (Glis glis) we show that age strongly affects hibernation/activity patterns, and that this occurs via two pathways: (i) with increasing age, dormice are more likely to reproduce, which delays the onset of hibernation, and (ii) age directly advances emergence from hibernation in spring. We conclude that hibernation has to be viewed not merely as an energy saving strategy under harsh climatic conditions, but as an age-affected life-history trait that is flexibly used to maximize fitness.

Figure 1

Path analysis of effects of age on reproduction on hibernation in Glis glis. The path analysis based on piecewise structured equation models for (a) males and (b) females. Boxes represent measured variables (red = age, blue = variables related to hibernation) and can appear as response variables in one path or as a predictor variable in another. For clarity, the binomial variable “sexually active” (Repro: yes/no) is also printed in one box. The impact of age on HibStart (hibernation onset) was partly indirect and caused by sexual activity. Arrows represent unidirectional relationships among variables (black = positive effect, red = negative effect). Shown are all included fixed variables, but only significant (P < 0.05) effects are represented by an arrow (for more details, i.e. non-significant effects see Supplementary information S1). The thickness of each significant path arrow has been scaled on the magnitude of the standardized regression coefficient. Asterisks indicate the significance level: *P < 0.05, **P < 0.01, ***P < 0.001, marginal and conditional R² are given in the text (see also Supplementary information S2). HibStart = onset of hibernation (day of year), HibDur = hibernation duration (d), HibEnd = end of hibernation (day of year), Age = log (age) in years, repro = sexually active yes/no (for details see Material and Methods), Individual quality = lifespan of the individual (years), Body mass was measured in g prior to hibernation.

Figure 2

Effect of age on the probability to reproduce in dormice (females: probability to give birth to a litter, males: probability to develop sexually competent testes in the current active season). Please note that, to ease visualisation, proportions (symbols), predictions (lines) and 95% confidence intervals (dashed lines) are shown, and all ages at or above 8 years were pooled due to small sample sizes. All statistical evaluations were based on individual data. For comparison the reproductive probability (females) or capability (males) as a function of age on a linear scale are shown in Fig. S4.1.

Figure 3

Partial effect of log (age) on weaned litter size in female edible dormice. Please note that these results were obtained from a subsample of our data set, for which the exact number of offspring of each female could be determined (N = 28 litters). For comparison, the partial regression plot of linear effects of age on weaned litter size is shown in Fig. S4.2.

Figure 4

Partial effects of age (log) measured in years on the end of hibernation in males and females. For details of the path analysis see text and Material and methods. For comparison, the partial regression plot of linear effects of age on the time of emergence from hibernation is shown in Fig. S4.3 (males) and Fig. S4.4 (females).

Figure 5

Age dependent yearly survival probabilities in enclosure housed edible dormice. Since there was no significant sex effect on survival, data for males and females were pooled. To ease visualisation, proportions (symbols), predictions (line), and 95% confidence intervals (dashed lines) are shown, and ages at or above 8 years are pooled due to small sample sizes. All statistical evaluations were based on individual data. The number of animals is given at the bottom of the figure. For comparison, effects of age (on a linear scale) on yearly survival in dormice kept in outdoor enclosures are shown in Fig. S4.5.

Figure 6

Graphical summary of effects of age and reproduction on the timing of hibernation and activity. As age increases both emergence from hibernation and hibernation onset occur earlier in the year, leading to a phase advancement of the active season (horizontal bars; blue: males; orange: females). Also, the proportion of sexually active males with large testes (blue triangles) and of reproducing females (yellow circles) increases with age. Investment into reproduction leads to a delay in hibernation onset in both sexes (light blue and yellow bars with arrows). Shown are model predictions for non-reproducing dormice, as well as predictions based on the observed fraction of reproductively competent/active animals in each age class.