Compared to other primates, childbirth is remarkably difficult in humans because the head of a human neonate is large relative to the birth-relevant dimensions of the maternal pelvis. In seems puzzling that females have not evolved wider pelvises despite the high maternal mortality and morbidity risk connected to childbirth in humans. In the proposed project, I will build a comprehensive novel model of the evolution of the human pelvis. With this project, I will contribute to answering the question why this ‘obstetric dilemma’ still exists in modern humans. Current models explaining the evolution of human pelvic dimensions, or the lack thereof, are based on verbal arguments only. In contrast, I propose to use empirical data on phenotypic variation of pelvic dimensions in human populations (data from populations in North America, Europe and Southern Africa) as well as mathematical models of evolutionary dynamics (quantitative genetic models) that build on these data to quantitatively assess hypotheses on pelvis evolution. In particular, I will estimate the phenotypic integration and evolvability of different pelvic shape features. In an earlier study, I have shown that evolution has produced covariances between pelvis shape and other body dimensions that contribute to ease childbirth. Based on these results, I will compare the variational properties of the pelvis across sexes and populations. I will devise a quantitative model of the constraints imposed upon the evolution of the human pelvis using the collected empirical data and data from the gynecological literature.