Living in groups can confer advantages, such as improved foraging, access to mates, or defence. It may also incur costs, for example through competition for resources. Eusocial species overcome this disadvantage of group living, but groups (colonies) vary in structure. One such structure is polydomy, found in ants, in which a single colony inhabits several nests. It has been hypothesised that polydomous colonies benefit from enhanced defences through:
(1) Dilution of risk: if one nest is destroyed, the colony will persist.
(2) Improved retaliation: inhabitants of a nest targeted by a predator may recruit from connected nests to aid in defence.
(3) Improved evacuation: the inhabitants of a nest have connected nests to flee to.
A dynamic network model of polydomy suggests that stochastic nest destruction favours polydomy over monodomy. This thesis adapts this model to investigate the effect of predator behaviours on the success of polydomy, and to investigate whether polydomous colonies could benefit from colony-level retaliation. We find that the frequency of polydomy in a population increases when a predator would destroy the entire nest, but that the addition of defence by retaliation has no effect on the frequency of polydomy. This suggests that polydomous colonies do benefit from dilution of risk, but not from improved retaliation. Extensions to address evacuation are discussed, as is the potential impact of including further predator behaviours, such as learning.
