The torso integration hypothesis revisited in Homo sapiens: contributions to the understanding of hominin body shape evolution

Objectives: Lower thoracic widths and curvatures track upper pelvic widths and iliac blades curvatures in hominins and other primates (torso integration hypothesis). However, recent studies suggest that sexual dimorphism could challenge this assumption in Homo sapiens. We test the torso integration hypothesis in two modern human populations, both considering and excluding the effect of sexual dimorphism. We further assess covariation patterns between different thoracic and pelvic levels, and we explore the allometric effects on torso shape variation. Material and Methods: A sex-balanced sample of 50 anatomically connected torsos (25 Mediterraneans, 25 Sub-Saharan Africans) was segmented from computed tomography scans. We compared the maximum medio-lateral width at seventh-ninth rib levels with pelvic bi-iliac breadth in males and females within both populations. We measured 1,030 (semi)landmarks on 3D torso models, and torso shape variation, mean size and shape comparisons, thoraco-pelvic covariation and allometric effects were quantified through 3D geometric morphometrics. Results: Females show narrow thoraces and wide pelves and males show wide thoraces and narrow pelves, although this trend is more evident in Mediterraneans than in Sub-Saharans. Equal thoracic and pelvic widths, depths and curvatures were found in absence of sexual dimorphism. The highest strength of covariation was found between the lowest rib levels and the ilia, and allometric analyses showed that smaller torsos were wider than larger torsos. Conclusions: This is the first study testing statistically the torso integration hypothesis in anatomically connected torsos. We propose a new and more complex torso integration model in H. sapiens with sexual dimorphism leading to different thoracic and pelvic widths and curvatures. These findings have important implications in hominin body shape reconstructions.