Predicting resource utilization: the utility of optimal foraging models

Predicting how predation changes communities is an important challenge for ecologists. One view assumes that predictions can be made in terms of behavioural phenomena. Static optimality methods have been used to devise models of prey choice and are thought to have performed well. Two examples from the literature show that models of diet choice have either been applied without first testing them at the individual level, or have been tested in a way that does not fulfil all their assumptions. A discussion of the nature of mathematical models shows why it is dangerous to translate theories into natural language and how important it is to appreciate the limitations on the mathematical functions describing behaviour. A final section discusses behavioural phenomena such as satiation and learning, which are likely to limit the predictive capacities of static optimization models. It is likely that dynamic models are the way forward although they may be harder to apply at the ecological level.