Biology and Philosophy 17: 551–565, 2002. © 2002 Kluwer Academic Publishers. Printed in the Netherlands. Article On Price’s Equation and Average Fitness BENJAMIN KERR ∗ Department of Biological Sciences Stanford University Stanford, CA 94305 USA PETER GODFREY-SMITH Philosophy Department Stanford University Stanford, CA 94305-2155 USA Abstract. A number of recent discussions have argued that George Price’s equation for rep- resenting evolutionary change is a powerful and illuminating tool, especially in the context of debates about multiple levels of selection. Our paper dissects Price’s equation in detail, and compares it to another statistical tool: the calculation and comparison of average fitnesses. The relations between Price’s equation and equations for evolutionary change using average fitness are closer than is sometimes supposed. The two approaches achieve a similar kind of statistical summary of one generation of change, and they achieve this via a similar loss of information about the underlying fitness structure. Key words: altruism, average fitness, clumping, group selection, individual selection, mar- ginal fitness, multi-level selection, Price’s equation, statistics 1. Introduction In the study of selection in structured populations, several researchers have suggested that Price’s (1970, 1972, 1995) famous equation is a powerful and illuminating statistical tool (Frank 1995; Hamilton 1975; Sober and Wilson 1998). What exactly is the Price equation and what does it really do? In this paper we try to answer this question by drawing on our companion paper (this volume). In addition, we discuss another well-known statistical approach – that involving average (or “marginal”) individual fitnesses. We will argue ∗ Current address: Ecology, Evolution and Behavior Department, University of Minnesota, 100 Ecology, 1987 Buford Circle, St. Paul, MN 55108, USA