Density-dependent effects on group size are sex-specific in a gregarious ungulate

Abstract

Density dependence can have marked effects on social behaviors such as group size. We tested whether changes in population density of a large herbivore (elk, Cervus canadensis) affected sex-specific group size and whether the response was density- or frequency-dependent. We quantified the probability and strength of changes in group sizes and dispersion as population density changed for each sex. We used group size data from a population of elk in Manitoba, Canada, that was experimentally reduced from 1.20 to 0.67 elk/km(2) between 2002 and 2009. Our results indicated that functional responses of group size to population density are sex-specific. Females showed a positive density-dependent response in group size at population densities ≥0.70 elk/km(2) and we found evidence for a minimum group size at population density ≤0.70 elk/km(2). Changes in male group size were also density-dependent; however, the strength of the relationship was lower than for females. Density dependence in male group size was predominantly a result of fusion of solitary males into larger groups, rather than fusion among existing groups. Our study revealed that density affects group size of a large herbivore differently between males and females, which has important implications for the benefits e.g., alleviating predation risk, and costs of social behaviors e.g., competition for resources and mates, and intra-specific pathogen transmission.

Figure 1. Relation between mean group size and population density for female (black) and male (gray) elk in Riding Mountain National Park (2002–2005 and 2007–2009).

Lines are quadratic fits to the mean group size data illustrating: (a) density-dependent change in females, including negative density-dependence at low density; and (b) weak density dependence for males.

Figure 2. Predicted change in unadjusted group size for female and male elk in Riding Mountain National Park (2002–2005 and 2007–2009); here the dichotomy between density-dependent response in group size for females and males is pronounced.

These models also control for seasonal sight-ability bias due to canopy cover (a) October – March, (b) April – September.

Figure 3. Predicted probability of encounter rate of groups (binned by size) for female (a) and male (b) elk with changes in population density in Riding Mountain National Park (2002–2005 and 2007–2009).

Figure 4. Predicted changes in group dispersion (binned by size) for female elk with changes in population density in Riding Mountain National Park (2002–2005 and 2007–2009).