The evolution of vertebral formulae in Hominoidea

Primate vertebral formulae have long been investigated because of their link to locomotor behavior and overall body plan. Knowledge of the ancestral vertebral formulae in the hominoid tree of life is necessary to interpret the pattern of evolution among apes, and to critically evaluate the morphological adaptations involved in the transition to hominin bipedalism. Though many evolutionary hypotheses have been proposed based on living and fossil species, the application of quantitative phylogenetic methods for thoroughly reconstructing ancestral vertebral formulae and formally testing patterns of vertebral evolution is lacking. To estimate the most probable scenarios of hominoid vertebral evolution, we utilized an iterative ancestral state reconstruction approach to determine likely ancestral vertebral counts in apes, humans, and other anthropoid out-groups. All available ape and hominin fossil taxa with an inferred regional vertebral count were included in the analysis. Sensitivity iterations were performed both by changing the phylogenetic position of fossil taxa with a contentious placement, and by changing the inferred number of vertebrae in taxa with uncertain morphology. Our ancestral state reconstruction results generally support a short-backed hypothesis of human evolution, with a Pan-Homo last common ancestor possessing a vertebral formulae of 7:13:4:6 (cervical:thoracic:lumbar:sacral). Our results indicate that an initial reduction in lumbar vertebral count and increase in sacral count is a synapomorphy of crown hominoids (supporting an intermediate-backed hypothesis for the origins of the great ape-human clade). Further reduction in lumbar count occurs independently in orangutans and African apes. Our results highlight the complexity and homoplastic nature of vertebral count evolution, and give little support to the long-backed hypothesis of human evolution.