Knowing the network of interactions between organisms is essential to understand ecosystem functioning. In particular, documenting food web structure allows to understand transfers of energy and matters. Among them, many pollutants, especially metallic trace elements, are transferred via food between organisms. Diet of animals and its spatial and temporal variations are however rarely taken into account to understand pollutants behavior in situ, likely because traditional diet analysis methods (histology and microscopy) are too time-consuming and not precise enough. The recent development of molecular methods is promising to evaluate diet diversity with higher precision and more quickly.
On a partially remediated former gold mine in southern France (Salsigne), we studied arsenic transfer in a food web involving small mammals, via a molecular assessment of both herbivorous and insectivorous diet. We especially measured the exposition of animals to arsenic (contamination of food) and the distribution of arsenic in organs, at 4 sessions of capture, at 2 seasons (spring and autumn) and on 4 levels of remediation. In parallel, we began to analyze the diet of the 4 species (wood mouse, Algerian mouse, common vole and greater white-toothed shrew) using the new metabarcoding approach on both stomach content and faeces, to detect animal preys and plant food items. We used the trnL method to analyse the plant diet of 10 small mammals. First results are very promising.