Individual behaviour, growth, survival and vulnerability to hunting in a large mammal

Abstract

Humans have exploited wild animals for thousands of years. Recent studies indicate that harvest-induced selection on life-history and morphological traits may lead to ecological and evolutionary changes. Less attention has been given to harvest-induced selection on behavioural traits, especially in terrestrial systems. We assessed in a wild population of large terrestrial mammals whether decades of hunting led to harvest-induced selection on trappability, a proxy of risk-taking behaviour. We investigated links between trappability, horn growth and survival across individuals in early life and quantified the correlations between early-life trappability and horn growth with availability to hunters and probability of being shot. We found positive among-individual correlations between early-life trappability and horn growth, early-life trappability and survival and early-life horn growth and survival. Faster growing individuals were more likely to be available to hunters and shot at a young age. We found no correlations between early-life trappability and availability to hunters or probability of being shot. Our results show that correlations between behaviour and growth can occur in wild terrestrial population but may be context dependent. This result highlights the difficulty in formulating general predictions about harvest-induced selection on behaviour, which can be affected by species ecology, harvesting regulations and harvesting methods used. Future studies should investigate mechanisms linking physiological, behavioural and morphological traits and how this effects harvest vulnerability to evaluate the potential for harvest to drive selection on behaviour in wild animal populations.

Keywords: animal personality; bighorn sheep; exploitation; harvest‐induced selection; hunting vulnerability; pace‐of‐life.

FIGURE 1

Population‐level effects of age and population size on bighorn sheep rams' trappability (a, b) and horn growth (c, d), at Ram Mountain, Alberta, Canada. The lines represent the model's predicted values of the responses for a given predictor when other predictors in the model are set to their mean value. The shaded areas around the lines represent the 95% uncertainty interval of the predicted responses. The distributions of the raw data are represented in the background, with the median indicated by the dots or the lines at the centre of the boxplots. Population size was grouped by tens to generate the raw data distributions.

FIGURE 2

Among‐individual differences in early‐life trappability (a) and horn growth (b) for bighorn sheep rams at Ram Mountain, Alberta, Canada. Repeatability estimates include variation due to fixed effects in the denominator while it was removed for adjusted repeatability estimates. Complete posterior distributions of estimates are illustrated. Shaded areas within the distributions represent the 95% credible intervals. Vertical lines represent the mean estimate.

FIGURE 3

Among‐individual correlations between early‐life trappability, horn growth and survival and availability to hunters and hunting fate at or prior to 6 years of age, for bighorn sheep rams at Ram Mountain, Alberta, Canada. Complete posterior distributions of estimates are illustrated whit horizontal lines representing the 95% credible intervals and dots representing the median estimates (a). Relationships between the posterior modes of the ‘Id’ random effect, that is, best linear unbiased predictions (BLUPs), are represented for each correlation for visualisation purposes (b–h).

FIGURE A1

Among‐year differences in early‐life trappability (a) and horn growth (b) for bighorn sheep rams at Ram Mountain, Alberta, Canada. Repeatability estimates include variation due to fixed effects in the denominator while it was removed for adjusted repeatability estimates. Complete posterior distributions of estimates are illustrated. Shaded areas within the distributions represent the 95% credible intervals. Vertical lines represent the mean estimate.

FIGURE A2

Among‐year correlations between early‐life trappability, horn growth and survival for bighorn sheep rams at Ram Mountain, Alberta, Canada. Complete posterior distributions of estimates are illustrated whit horizontal lines representing the 95% credible intervals and dots representing the mean estimates (a). Relationships between the posterior modes of the ‘Id’ random effect, that is, best linear unbiased predictions (BLUPs), are represented for each correlation for visualisation purposes (b–d).

FIGURE A3

Variability of trappability with age of male bighorn sheep at Ram Mountain, Alberta, Canada. (a) Distribution of trappability for all ages. (b) Variance in trappability per age. The size of the dots and the number next to them reflect the sample size used to calculate the variance, that is, the number of individuals of that age. (c) Population‐level effect of age on trappability. The line represents the model's predicted values of trappability with age, at mean population size. The shaded area around the line represents the 95% uncertainty interval of the predicted response. The distribution of the raw data is represented in the background. (d) Individual trajectories of trappability with age. Grey lines represent individual Poisson regressions of trappability as a function of age using the raw data. The shaded area represents the overlapping 95% prediction intervals of the model for each individual accounting for population size and random intercepts and slopes of individual identity.