Covariation between human pelvis shape, stature, and head size alleviates the obstetric dilemma

Abstract

Compared with 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. It seems puzzling that females have not evolved wider pelvises despite the high maternal mortality and morbidity risk connected to childbirth. Despite this seeming lack of change in average pelvic morphology, we show that humans have evolved a complex link between pelvis shape, stature, and head circumference that was not recognized before. The identified covariance patterns contribute to ameliorate the "obstetric dilemma." Females with a large head, who are likely to give birth to neonates with a large head, possess birth canals that are shaped to better accommodate large-headed neonates. Short females with an increased risk of cephalopelvic mismatch possess a rounder inlet, which is beneficial for obstetrics. We suggest that these covariances have evolved by the strong correlational selection resulting from childbirth. Although males are not subject to obstetric selection, they also show part of these association patterns, indicating a genetic-developmental origin of integration.

Keywords: correlational selection; evolution; morphometrics; obstetric dilemma; pelvis.

Fig. 1.

Pelvic landmarks. The full set of 126 3D pelvic landmarks measured on each pelvis, shown as red spheres on the mean pelvis shape, are shown in (A) anterior, (B) superior, and (C) lateral view. The mean pelvis shape was computed as the average shape of all individuals in our dataset.

Fig. 2.

Sexual dimorphism in the human pelvis. (A) The average female pelvis shape and (B) the average male pelvis shape in the sample, in frontal view. The differences between these two average shapes illustrate well-known patterns of sexual dimorphism in the human pelvis. Females have a broader and flatter pelvis, a wider and shallower pelvic cavity, a wider subpubic angle, and smaller acetabula than males.

Fig. 3.

Association between pelvis shape and stature, illustrated by average pelvis shapes for individuals with short and tall stature, separately for females and males. The shape differences shown here correspond to the partial linear regression coefficients for stature from the shape regressions. Hence, they represent the association of pelvis shape with stature, independent of head circumference. Each of these pelvis shapes is shown in anterior, superior, and lateral view (Top, Middle, and Bottom, respectively). The magnitude of the displayed shape differences corresponds to a deviation of ±40 cm in stature from the sample average, which is approximately a twofold extrapolation of the actually occurring variation. On average, taller persons have a taller and narrower pelvis with longer ilial blades and a shorter relative distance between the acetabula compared with shorter persons. Taller persons also have a more oval pelvic cavity with an outward-projecting pubic symphysis, whereas short persons have a rounder pelvic cavity. The relative height of the sacrum and the symphysis increases with stature in males. This effect is weakly present in females.

Fig. 4.

Association between pelvis shape and head circumference, illustrated by average pelvis shapes for individuals with small and large head circumference, separately for females and males. These shape differences correspond to the partial linear regression coefficients for head circumference from the shape regressions. Hence, they represent the association of pelvis shape with head circumference, independent of stature. Each of these pelvis shapes is shown in anterior, superior, and lateral view (Top, Middle, and Bottom, respectively). The magnitude of the displayed shape differences corresponds to a deviation of ±10 cm in head circumference from the sample average, which is approximately a twofold extrapolation of the actually occurring variation. Both in males and females, a round pelvic cavity is associated with a large head, whereas an oval pelvic cavity is associated with a small head. On average, females—but not males—with a large head have a shortened sacrum projecting outward from the birth canal.

Fig. 5.

Shape regression scores. Shown is a scatterplot of stature versus the regression scores from the regression of pelvic shape on stature (with head circumference as covariate). As these scores are projections of the individual pelvis shapes on the corresponding vector of regression coefficients (linear combination of the shape variables weighted by their covariance with stature), they are the shape scores with maximum covariance with stature. The scatterplots are computed separately for (A) females and (B) males. Similar scatterplots of head circumference versus the corresponding regressions scores are shown in C and D for females and males. Open circles represent female pelvises, and filled circles represent male pelvises.