Evolution in the palm of the human hand : functional inferences from internal bone architecture in great apes and fossil hominins

When and how did arboreal environments become less important, and technology become more important, for hominin survival? This fundamental question of our own evolution has been a focus of palaeoanthropological and archaeological debate for over a century. The present work aims to address this scientific problem via the analysis of the internal structure of osteological and fossil material. Specifically the trabecular, or cancellous, epiphyseal structure and the cortical diaphyseal structure of long bones, have been experimentally demonstrated to change in response to loads they experience during life. This process, known as 'bone functional adaptation' (Cowin et al., 1985; Ruff et al., 2006), holds that bone will functionally adapt to its mechanical environment and so its architecture may be used to infer the behaviours that are consistent with this environment. Therefore, this internal bone architecture can be used to infer loads experienced by fossil hominin hands and thus hand postures used in arboreal locomotion or manipulation, consistent with these loads. The following analyses focus on metacarpals because they are the long bones of the hand, amenable to diaphyseal and trabecular analysis. The metacarpophalangeal joints are bi-axial, while the trapeziometacarpal joint is further mobile, and together they provide a range of potentially loaded hand postures adopted in different grips. The internal bone architecture of fossil and osteological hand bones was analysed via micro-computed tomography. Chapter 1 - General Background. Chapter 1 contextualises the scientific problem within an in-depth discussion of the manual fossil record and bone functional adaptation. It describes in detail why functional inference based on external fossil morphology alone is open to interpretative debate. Chapter 2 - Materials and Methods. Chapter 2 describes the methods employed, from a hand bone on the collection bench to an analysed, segmented virtual model. It also provides a rationale for each successive stage of analysis. Chapter 3 - MIA-Clustering: A novel method for segmentation of paleontological material. While mentioned in Chapter 2, this chapter presents the detailed description and testing of a new method of image segmentation, developed for the present thesis, in order to accurately delineate trabecular morphology in fossils: MIA-Clustering. Chapter 4 - Metacarpal trabecular bone varies with distinct hand-positions used in hominid locomotion. Chapter 4 validates the assumption that the trabecular structure of the second to fifth metacarpals of extant hominids is consistent with what is known about hand loading, during the diverse habitual locomotor repertoires, of hominids. It provides validation of the method developed to test for inter-specific differences in trabecular distribution and, unexpectedly, provides a method of informal falsification for the current mode of fossil hominin behavioural inference. Chapter 5 - Trabecular variation in the first metacarpal reflects distinctive human manipulation among hominids. Chapter 5 validates the assumption that the trabecular structure of the first metacarpal of extant hominids is consistent with the observed manipulative grips habitually used by great apes, including modern humans. Chapter 6 - Fossil hominin hand use: Evidence for arboreality and human-like manipulation in Australopithecus. This chapter infers the likely grips of fossil hominins from both their trabecular and cortical structure. Importantly, it brings together all of the rays in a holistic analysis of internal fossil morphology, in the context of internal bone structure consistent with arboreal and manipulative grips practised by extant hominids. Chapter 7 - Discussion and Conclusion. This chapter discusses how far the present results resolve the initial scientific problem and what we can say about fossil hominin hand use. Further this chapter acknowledges the limitations of the current approach and what future data or methodologies may ameliorate them.