Résumé
Since 2016, we are currently experimenting with underground silos for storing old varieties of organically grown cereals, under a Mediterranean climate by using hand tools. Our objectives are to re-discover ancient knowhow and shed new lights on the pits for helping archaeologist to interpret them. This original decade-long program combines the expertise of various disciplines (archaeology, carpology, microbiology) with that of actors of the cereal food chain (farmers, millers, bakers). We are seeking to identify the parameters that lead to the preservation of grain placed in an underground silo, to assess the stability of the device for a single or repeated use, and to evaluate the relevance of long-term preservation (several months or several years), depending on the use of the grain (for trade, subsistence or sowing). In archaeology, the paucity of evidence of the stored product in favour of the omnipresence of the container (the pit), perhaps makes us forget that the shape, volume and elements resulting from the closure or development of the walls, are not only cultural or chronological criteria but also the witnesses of an agricultural technique specific to an environment.Our experiments focus on the underground silo as an air tight unit for grain storage. The increase in the number of pits could provide food for thought about the management of past silage areas, and, in particular, the large ones of early medieval southern France. Every stage, from the digging of the silo to the removal and carbonisation of the grain in the laboratory, is tested. The intervention schedule is as close to known agricultural practices as possible, and adapted to the stored crop cycle. This multi-year programme brought together an interdisciplinary team, inspired by François Sigaut’s approach presented at the conference on grain conservation techniques held in the early 1980s (Gast and Sigaut, 1979). Initiated by archaeologists, this experimental programme was a response to the struggle of correctly identifying past ensilage techniques on archaeological sites and, above all, to understand the conservation process in a confined underground atmosphere without the product in place. Indeed, archaeological studies almost exclusively focus on the container – the pit – which is often backfilled after being used for storage, then dismantled over the centuries. The gaps in the archaeological data are partly offset by the occasional discovery of exceptionally well-preserved pits, in a state of preservation close to the original silo, and by the valuable contributions of carpology. But, even in the most promising cases, the analysis still leaves some grey areas. The charred plant remains encountered in archaeological silos are not necessarily direct evidence of the ensiled product, but rather waste products (Ros et al. 2019). Regarding the closing devices or internal protection of the grain (at walls level), this is generally of variable interest and deduced from stratigraphic analysis. Furthermore, without understanding the context and purpose of silage-making, it’s difficult to understand why one type of technique was chosen over another. Archaeology still lacks another vital piece of information: how long can food products be stored. We therefore turn to other sources for explanations (agronomic texts, ethnographic accounts, other archaeological experiments), but these examples are often from contexts too far removed from the reality of the discoveries, or too theoretical.These various observations explain our decision to practise traditional silage making under a Mediterranean climate, in an area where sites from all periods are plentiful. In the Pyrénées-Orientales (France), many archaeological silos have been discovered, isolated or grouped together, at times associated with a dwelling or other storage facilities. On the basis of experiments carried out in 2018 and 2019, the aim of this paper is to discuss a specific point commonly accepted by archaeologists: if opened, the silo must be completely emptied of its food content, otherwise the sudden supply of oxygen would irremediably interrupt the conservation process. Our storage results, analysed by grain specialists, are helping to revise this assumption. We now consider that opening the silo to manipulate the stock during the process is possible and does not hinder the conservation process. Although this programme spreads over to 2026, not allowing yet to grasp all the issues involved in successful conservation, this practice is already helping to revise this paradigm.