Dans une démarche pluridisciplinaire, qui emprunte aussi bien aux sciences humaines et sociales (archéologie, histoire des techniques, réexamen des textes anciens) qu’aux sciences chimiques (science des matériaux, métallurgie expérimentale, chimie analytique), cette recherche vise à reconstituer l’évolution des techniques de fonte à la cire perdue, depuis les plus anciens témoignages de son utilisation à Mehrgarh (Pakistan, 5e millénaire av. J.-C.) jusqu’aux premières grandes statues de bronze dans le monde égéen (fin du VIe, début du Ve s. av. J.-C.).L’étude des objets a parfois nécessité des développements analytiques spécifiques pour pallier à leur très forte altération, comme par exemple l’imagerie de photoluminescence synchrotron. Cela a notamment permis de reconstituer en détail la chaîne opératoire de fabrication de la rouelle de Mehrgarh, l’une des plus anciennes fontes à la cire perdue connue à ce jour. Sur la base de ces résultats, il est suggéré que la fonte à la cire perdue pourrait avoir été inventée pour donner la possibilité aux individus non métallurgistes de créer des objets importants en métal tels que ces amulettes, par le biais du façonnage d’un modèle en cire.Nous montrons que la cire perdue a ensuite été mise à profit pour donner naissance à une nouvelle forme de sculpture, la statuaire de métal. Au prix de parois très épaisses et d’assemblages mécaniques, nous mettons en évidence une première période de production de grandes statues en Mésopotamie pendant la deuxième moitié du 3e millénaire av. J.-C. Après un long hiatus, la grande statuaire métallique renaît de façon spectaculaire pendant la première moitié du 1er millénaire av. J. C., à la fois dans les mondes égyptien, sabéen et égéen. Nous identifions deux innovations importantes responsables de cette renaissance : le procédé indirect et l’assemblage soudé.Des essais de coulabilité effectués en faisant varier la composition de l’alliage et le matériau du moule sont présentés dans la dernière partie de notre travail. En moule de plâtre et avec un fort préchauffage, une coulabilité exceptionnelle a été obtenue pour le bronze à fort taux de plomb, donnant pour la première fois une idée des conditions de coulée nécessaires à l’obtention des parois très minces souvent observées sur les grandes statues antiques. / Following a multidisciplinary approach combining social sciences (archaeology, history of techniques, examination of ancient texts) and chemistry (materials science, experimental metallurgy, analytical chemistry), our research aims to reconstruct the evolution of lost-wax casting techniques, from the earliest evidence of its use in Mehrgarh (Pakistan, 5th millennium BC) to the first large bronze statues in the Aegean (end of the 6th, beginning of the 5th century BC).The archaeological artefacts under study have sometimes required specific analytical developments to overcome their very strong alteration, such as synchrotron photoluminescence imaging. This allowed reconstructing in detail the chaîne opératoire of the Mehrgarh wheel-shaped amulet, one of the oldest lost-wax castings known to date. Based on the results thus obtained, it has been suggested that lost-wax casting might have been invented to give non-metalworkers the opportunity to create important metal objects such as amulets by simply shaping a wax model.We show that lost-wax casting was afterwards used to create a new form of sculpture, namely metal statuary. At the cost of very thick metal walls and mechanical assemblies, this led to an early production step of large statues in Mesopotamia during the second half of the 3rd millennium BC. After a long hiatus, large metal statuary reappeared spectacularly during the first half of the 1st millennium BC, both in the Egyptian, the Sabean and the Aegean areas. We were able to identify two important innovations responsible for this renaissance: the indirect process and the flow fusion welding.Castability tests carried out by varying the composition of the alloy and the material of the mould are presented in the last part of our work. An exceptional castability was obtained for highly-leaded bronze in plaster mould and with a high preheating. For the first time some insights are thus drawn on the casting conditions necessary to obtain the very thin walls often observed on large metal statues during the Classical Antiquity.

This thesis investigates the relationship between climate, agriculture and social change in South Asia’s Bronze Age urban Indus Civilisation. Specifically, my research tests the hypothesis that an abrupt weakening of the Indian Summer Monsoon ca 2100 cal BC led to increasing crop water stress, and hence potentially contributed to the Civilisation’s decline by reducing food supply. This hypothesis is frequently invoked in discussions of the Civilisation’s end, yet until now, has not been empirically tested. Using material excavated from several Indus settlements, this study uses a novel combination of isotopic techniques to directly test the connection between climate change and agricultural stress. These techniques are first, oxygen isotope analysis of faunal bones and teeth; and second, stable carbon isotope analysis of crop remains. The oxygen analyses provide detailed records of monsoon intensity at a local, human scale, while the carbon analyses provide an empirical test of whether crop water stress increased. Applied in parallel across a diverse suite of Indus sites, these techniques together provide an archaeologically and ecologically-nuanced analysis of climatic impacts. The archaeological analyses are supported by a methodological study, which investigates how water status relates to the stable carbon isotope signature in barley (Hordeum vulgare) and the Indian jujube (Ziziphus mauritiana) along a climatic transect in north-western India today. Overall, the isotopic results suggest that at the sites sampled here, climate change probably had minimal impacts on crop water availability. This does not necessarily mean that climate change had no impacts on agriculture across the greater Indus sphere, and indeed there are hints that there may have been climatic stress in more vulnerable settings. However, at the sites studied here, any hydrological consequences of climate change—including the 4.2 ka event—appear to have had neither a lasting nor a pervasive impact on the adequacy of crop water supply. This is an important finding, and necessitates a clear refinement of how we think about climatic sensitivity, climatic vulnerability, and climatic impacts across—and indeed beyond—the greater Indus.