Behavioral estimates of male mating success tested by DNA fingerprinting in a polygynous mammal

We used DNA fingerprinting to determine paternity of 80 calves born into the individually monitored population of red deer (Cervus elaphus) on the Isle of Rum, Scotland. Using the paternity information, we investigated the assumptions and predictions of behavioral estimates of male mating success, on which previous studies in this population have relied. Over an 11- day interval in the rut centered on a female's estimated date of conception, the probability that a male fathers a calf is closely related to the number of days he holds the female in his harem, increasing from 0.12 for 1 day to 1.0 for 6 days or more. We compared three methods for estimating the reproductive success of individual males from field observations of harem membership with true success revealed by DNA fingerprinting. All three methods accurately identified the relative success of individual males but were poor predictors of absolute success: the behavioral methods underestimate the true success of successful males and overestimate the success of many males who, in fact, fail to father any calves. In consequence, variance in male mating success is greater than previous behavioral estimates for this population suggested. Both harem membership data and observations of mating and other estrous behavior can be used to identify males most likely to father a specific calf.