Alleles at one gene are generally associated with alleles at many other genes (i.e., individuals carrying allele A of gene 1 are more likely to carry allele B of gene 2 than individuals carrying a). These associations arise because of functional interactions among genes, population subdivision, and simply as a necessary consequence of finite population size. One particularly important form is when beneficial alleles are negatively associated with each other, interfering with selection for each other. How strong is this effect in natural populations, and how can we detect it? How much recombination is needed to break down interference?