The authors adopted a new and attractive approach, based on the use of thermal springs appearing at different times, to study the short-term and long-term (> 50 years) effect of warming on the soil C stock under subarctic grasslands. This new approach allows to take a new look at the question of a positive feedback between temperature and soils that can amplify global warming. Indeed, most studies on this subject are based on warming experiments conducted over the short term (some years) or on questionable correlative approaches where the temperature co-varies with many other factors (e.g., study of soil C stocks along latitudinal temperature gradients). Their study challenges the current dominant view on the effect of warming on the dynamics of SOM. Indeed, results suggest that soil C losses in the subarctic grasslands studied cease after 5 years of warming. These observations corroborate those obtained in the rare ecosystem warming experiments maintained beyond 10 years. In addition, results suggest that the C stocks present in the deep soil horizons, where plant roots are not or hardly present, are not affected by warming. These unexpected discoveries, together with other recent observations, show the glaring lack of knowledge on the fundamental mechanisms of the effect of temperature on catalytic processes, which seriously compromises our ability to predict the soil-climate feedback.