Abstract
Human marriage systems, characterized by long-term partnerships and extended windows of parental care, differ from the mating systems of pulsed or seasonally breeding non-human animals in which Bateman's principles were originally tested. These features, paradigmatic of but not unique to humans, complicate the accurate measurement of mating success in evaluating Bateman's three principles. Here, we unpack the concept of mating success into distinct components: number of partners, number of years partnered, the timing of partnerships, and the quality of partners. Drawing on longitudinal records of marriage and reproduction collected in a natural-fertility East African population over a 20-year period, we test and compare various models of the relationship between mating success and reproductive success (RS), and show that an accurate assessment of male and female reproductive behaviour requires consideration of all major components of mating success. Furthermore, we demonstrate that while Bateman's third principle holds when mating success is defined in terms of years married, women's fitness increases whereas men's fitness decreases from an increase in the number of marriage partners, holding constant the total effective duration of marriages. We discuss these findings in terms of the distinct, sex-specific pathways through which RS can be optimized, and comment on the contribution of this approach to the broader study of sexual selection.