I. F. Harvey

Sperm precedence patterns are typically highly variable within (and between) species. Intraspecific variation in sperm precedence (measured as P₂, the proportion of progeny fathered by the last male to mate) is frequently seen as a candidate for adaptive interpretation through either male effects (e.g. body size), female effects (e.g. cryptic female choice) or an interaction between the two. Here we show, using computer simulation, that if ejaculates divide into a number of ‘packets’ and packets from two males mix randomly, then a variety of patterns of sperm precedence may result. We term this process ‘sloppy’ mixing. If ejaculates break into a small number of packets, bimodal P₂distributions are predicted. As the number of packets is increased, then a complex series of changes through multimodal and flat to unimodal distributions results. Sloppy mixing can thus result in many of the observed P₂ distributions. Sloppy mixing is unlikely to change the predictions of adaptive models of sperm competition.