A Learning-Based Model of Territory Establishment

Despite widespread interest in territorial behavior, the processes by which animals establish territories are still poorly understood. We present a new learning-based model of territory establishment for species in which individuals set up territories within large patches of spatially heterogeneous habitat. The model is based on the simple assumptions that individuals tend to return to areas in which they previously had rewarding experiences and, conversely, tend to avoid areas in which they previously engaged in costly aggressive interactions. The literature on learning and territorial establishment suggests that these assumptions are probably valid for many animals. Individual-based, spatially-explicit simulations of settlement behavior incorporating the assumptions of this model generate a number of phenomena comparable to those observed in territorial animals, including the formation of stable home ranges within large patches of uniform quality habitat, increases in territory size and home range exclusivity if settlers interact aggressively with one another, greater benefits of aggressive behavior if individuals settle at high density than if they settle at low density, higher success for residents when they compete with newcomers for the same space (the "prior residency advantage"), and the avoidance by newcomers of areas used by previous residents. Although the model needs further refinement to generate some phenomena observed in territorial species, our results suggest that the processes responsible for generating several basic components of territorial behavior may be simpler than is currently supposed.