Introduction

During the 1960s, a number of separate fields of research into the general question of the adaptive significance of behaviour could be identified. Within the traditions of classical ethology, people were conducting simple, direct experimental studies of the survival value of behaviour (as pioneered by Tinbergen, 1951). For example, Krebs et al., 1972 demonstrated experimentally that flocking in birds confers foraging advantages, and Neill and Cullen (1974) did the same for the anti-predator advantages of schooling in fishes. At the same time, in a rather different context, an increasing number of comparative studies, especially in birds (Lack, 1968; Crook, 1965) and primates (Crook and Gartlan, 1966), were beginning to show how social organisation can be related to ecological factors such as predation risk and food distribution. Also, people were seeking explanations for apparently altruistic behaviour, such as sacrificing personal reproductive output in favour of another individual or using ritualised aggression when fierce fighting would carry the day. Dissatisfaction with explanations couched in terms of group selection (WynneEdwards, 1962) led to recognition of the importance of inclusive fitness, and the development of the theory of kin selection (Williams, 1966; Hamilton, 1964; Maynard Smith, 1964, and see Dawkins, 1976). In the context of aggression, games theory analyses were used by Maynard Smith and collaborators (Maynard Smith and Price, 1973; Maynard Smith and Parker, 1976) to show how frequencydependent costs and benefits are sufficient to explain the existence of ritualised fighting. This was a special case of the economic approach pioneered by MacArthur (e.g. MacArthur and Pianka, 1966), whereby models incorporating the costs and benefits of behaviour were used to identify behaviour patterns that maximise fitness and so to elucidate the ways in which natural selection acts on behavioural traits.