INTRODUCTION

In a variety of mammals and a few birds, newly immigrated or newly dominant males are known to attack and kill dependent infants (Hausfater & Hrdy, 1984; Parmigiani & vom Saal, 1994; van Schaik & Janson, 2000). Hrdy (1974) was the first to suggest that this bizarre behaviour was the product of sexual selection: by killing infants they did not sire, these males advanced the timing of the mother's next oestrus and, owing to their new social position, would have a reasonable probability of siring this female's next infant. Infanticide would therefore be one of the most dramatic expressions of sexual conflict (Smuts & Smuts, 1993; Gowaty, 1997, this volume).

Although this interpretation, and indeed the phenomenon itself, has been hotly debated for decades (e.g. Dolhinow, 1977; Boggess, 1984; Bartlett et al., 1993; Sussman et al., 1995), on balance, this hypothesis provides a far better fit with the observations on primates than any of the alternatives (cf. van Schaik, 2000a). First, several detailed studies showed that the males never attacked or killed their own offspring (Borries et al., 1999; Soltis et al., 2000), in accordance with the more anecdotal information compiled from all directly observed cases of infanticide in the wild (van Schaik, 2000a). Second, several large-scale studies have estimated that the time gained by the infanticidal male amounts to 25 per cent, 26 per cent and 32 per cent of the mean interbirth interval (Crockett & Sekulic, 1984; Sommer, 1994; Borries, 1997).