We measured the energy cost of mate sampling by female pronghorn (Antilocapra americana), a species for which there are no apparent direct benefits of mate choice and for which the sampling tactic most closely resembles best-of-n or comparative Bayes. We used Global Positioning System collars to record the position of individuals at 10-min intervals during the 2 weeks preceding estrus in females that actively sampled and in females that did not sample. The difference in the 2-week energy costs of these two classes of females was 8,200 (+/-2,300) kJ, or roughly one-half of the energy cost of a single day. This value, expressed as the fraction of total yearly energy expenditure, is 59 times the value reported for a lekking bird. Our finding calls into question the common assumption in models of mate search that the cost of search is negligible as well as the common assumption that the cost of sampling must be small when there are only indirect benefits of female choice.