It is well known that soil is subject to many natural and human-induced threats, which contribute to its degradation and loss. In addition, sampling for scientific studies and professional activities increase soil loss, because the samples - often in excessive amount - are never restored to the original place. Here, we propose to call “hidden erosion” this unwelcome form of erosion. A rough estimation of the "hidden erosion" has not been carried out before. In this paper, the amounts of soil sampled from its original site and transferred to the laboratory is estimated using data of selected literature studies; the equivalent soil thickness reduction is calculated. From these estimates, we demonstrate that, in most cases, the amounts of soil sampled exceed the tolerable loss. This erosion form cannot usually heavily impact the environment, because of its limited spatial and temporal extent. However, when the soil is removed repeatedly on a large spatial scale, sampling may increase the erosion rates of the natural and other human-induced processes at the global scale. Finally, since sampling cannot usually be avoided in soil studies, a decalogue with some careful measures to minimize the soil loss during sampling is proposed, in order to limit the "hidden erosion".
