INTRODUCTION

Sexual dimorphism in anthropoid primates is commonly viewed as a product of sexual selection (Clutton-Brock et al., 1977; Harvey et al., 1978; Gaulin & Sailer, 1984; Clutton-Brock, 1985; Milton, 1985; Rodman & Mitani, 1987; Kay et al., 1988; Ely & Kurland, 1989; Greenfield, 1992a, b; Plavcan & van Schaik, 1992, 1994, 1997; Ford, 1994; Martin et al., 1994; Mitani et al., 1996b; Lindenfors & Tullberg, 1998; Plavcan, 1999, 2001; Barton, 2000; Lindenfors, 2002a; Mitani et al., 2002). Yet dimorphism in anthropoids is highly variable, and is expressed not as a single character, but rather to different degrees in different traits. This naturally raises the question of whether this variation is owing to variation in the strength of sexual selection, phylogenetic effects or the action of other selective factors on the dimorphic characters. While numerous papers have examined the causes and correlates of dimorphism in anthropoids, the relative contribution of sexual selection and other factors to variation in dimorphism remains unclear.

Part of this problem lies in the way that both dimorphism and sexual selection are measured. Both of these variables are estimated with error, not only in a simple statistical sense, but also in the assumptions that are used to justify measures as appropriate for analysis. Some of these biases are obvious, while others are not. Thus, if dimorphism is poorly correlated with an estimate of sexual selection, we can legitimately ask whether dimorphism is affected by factors other than sexual selection, or whether our measures fail to capture variation in either sexual selection or the targets of sexual selection.