THE EVOLUTION OF HELPING-BEHAVIOR IN LARGE, RANDOMLY MIXED POPULATIONS

A model is presented in which individuals who perform helping behavior can increase in frequency from low initial levels in a large population within which interactions take place between randomly selected individuals. This is accomplished without requiring extreme conditions. The model uses a payoff matrix that determines the consequences of each possible type of interaction between individuals. This payoff matrix changes as a function of the genetic constitution of the population. As a result, the payoff matrix may or may not satisfy the inequalities that define the prisoner''s dilemma, a formal game commonly used to study the evolution of helping behavior. An analysis of the model reveals that, for a wide range of parameter values, all stable equilibria occur at points where the genetic constitution of the population allows for satisfaction of the prisoner''s-dilemma inequalities. This is the case even though these inequalities may not be satisfied during the initial stages of invasions by helpers.