Understanding and predicting the residence time of soil organic carbon is a major issue in the context of climate change. A fraction of soil organic matter persists over decades and centuries, but no method allows, so far, its quantification and isolation. We proposed to use bare fallows plots, kept free of plants for decades, in which almost all C inputs have been stopped. They provide a unique opportunity to monitor the decay of soil organic C by mineralization under natural conditions and to quantify and isolate stable soil C. We set a network of long-term bare fallows (>30 yrs), covering a range of soil types and climate conditions, located at Askov (Denmark), Grignon and Versailles (France), Kursk (Russia), Rothamsted (UK) and Ultuna (Sweden). Using freshly sampled or archived soil samples, we quantified the centennial scale stable organic carbon pool and investigated the processes responsible of its persistence. We found in particular that persistent OM has a low energy content. Using this exceptional set of archived samples we demonstrated that the mineralisation of persistent organic carbon is more sensitive to temperature than labile organic carbon pools, adding strong arguments to a lively debate in the scientific community. Bare fallows, usually references in long-term agricultural experiments, are powerful tools to study soil organic matter dynamics, with a special emphasis on decadal and centennial scale persistent carbon pools. Long-term agricultural experiments may then allow to answer scientific questions for which they were not initially created for, which adds to their value for the scientific community.