Sexual selection theory predicts that sex roles will be determined by the operational sex ratio (OSR), the sex ratio among individuals searching for mates at any given time. There are two predictions: (i) the sex which is in 'excess' will be the more competitive sex with respect to access to mates; and (ii) the sex of which there is a 'shortage' will be the more choosy with respect to potential partners. We examine the second prediction and find that current OSR theory does not consider an important factor which affects mate choice. This factor is sex differences in variation in mate quality. Hence, we develop a new model of mate choice which shows that the parameter which should be optimized during mate choice is the trade-off between reproductive rate and mate quality. If mate choice is too lax, reproductive rate may be high but partners will be of low quality. If mate choice is too stringent, partners will be of high quality but reproductive rate will be low because such partners will be rare. Stringency of mate choice is, therefore, a facet of OSR theory. Indeed, our model shows that OSR theory can be used to integrate the effect of sex differences in both mating rate and variation in mate quality to predict the direction of mate choice. Our model suggests that: (i) mate choice is only selected when individuals of the opposite sex vary in their quality as mates; (ii) if the extent of variation in mate quality is equal within each sex, the sex with the lower potential mating rate will be the more choosy sex; but (iii) if there is sufficiently greater variation in mate quality among the sex with the lower potential reproductive rate, the sex with the higher potential mating rate will be the more choosy sex.(ABSTRACT TRUNCATED AT 250 WORDS)