Abstract. At the National Bison Range (western Montana, U.S.A.) mate choice by female pronghorn, Antilocapra americana, is clearly observable. Females move independently, and copulate once per oestrus. Males cannot force copulation, only temporarily block female movement, and do not monopolize resources critical to females. Females practice three distinct strategies of mate choice. 'Sampling' females visit several harem-holding males, remain with each male a short time, and switch between males at an increased rate as oestrus approaches. Switches often appear to be energetically expensive. The actual mating visit is brief, about 1·5 days. Sampling females always leave males that fail to defend an adequate zone of tranquillity around the harem, but also leave males when there is no apparent cause. The majority (71%) of exits from males are of the latter type. Most sampling females return to and mate with a male that they visited within the week before oestrus. 'Inciting' females behave as samplers until oestrus; then they move away from the harem male, inciting fights and other aggressive competitition. Inciting females watch the competition, and they always mate immediately with the winning male. Finally, a small proportion of females each year practise a 'quiet' strategy. More than a month before oestrus, they move to an isolated, peripheral location occupied by a single male, and remain with that male throughout oestrus. Some evidence suggests that the quiet strategy may be used when females have low energy reserves. Among samplers, there is no evidence that females copy mating decisions of others. Mating success of males was not correlated with body size, horn size, age, rate of display, or with activity budgets of females in harems they controlled. Male mating success was strongly correlated with haremdays, the sum of each day's harem size across the rut. Because harem defence demands sustained, often strenuous activity by males, the behaviour of females is a direct assay of male vigour, and therefore a likely case of mate selection by a good genes criterion.