Group selection and the evolution of altruism

'Group selection' is an oft-cited but controversial explanation for the survival of altruism. Rather than enter this debate, this paper asks whether a group structure alone can provide a reasonable explanation for the survival of altruistic behaviour in an evolving population. If altruism (defined as taking a cooperative action in a standard Prisoners' Dilemma) is to flourish, either groups need to be isolated from one another for multiple generations, or groups need to be constructed in a positively-assortative manner. Concentrating on the former case the paper shows that the size of the group, the relative benefit to cost of altruism and the number of generations in isolation all play a critical role. The smaller the group size, or the larger the benefit-to-cost ratio, the higher the survival chances of altruism. Additionally, for altruism to survive, the number of generations spent in isolated groups must be neither too big nor too small. Copyright 2004, Oxford University Press.