Abstract
Feeding behavior is described from an evolutionary perspective, and implications for modern neurobiological studies are suggested. In particular, it is argued that meals may have evolved more for sociocultural reasons than physiological imperatives, and that biological approaches to the study of feeding episodes should adopt a more flexible model that is founded in economic or cost-benefit considerations. Specific examples of flexibility in mouse feeding behavior are given. It is further argued that the modern human food environment is so immoderate that physiological manipulations designed to restrain eating have little hope of achieving this goal.