Surface waters of the Ardières and Morcille Rivers (South-East France) are historically polluted by arsenic (As), with total dissolved As concentrations reaching up 40 µg/L. High frequency monitoring of surface waters during 3 years showed a strong seasonal variability of total dissolved concentrations of As, with low concentrations in winter (high-water level, 2-3 µg/L) and high concentrations in summer (low-water level, 20-40 µg/L). These findings suggest a more pronounced input of low- oxygenated water from groundwater to the river during summer than during winter period. Under this assumption, we suggest that during summer, dissolved concentrations of reduced As (As III) are probably higher than during winter period. Since the reduced form of As (As III) is known to be more toxic than the oxidized form of As (As V) for aquatic organisms, there is a need to assess temporal variability of inorganic As speciation in these surface waters. Nevertheless, inorganic forms of As are unstable, especially when parameters such as pH and/or redox vary. Passive sampling technique such as Diffusive Gradient in Thin films (DGT) presents several advantages to better assess in-situ As speciation, by eliminating problems of contamination and especially by reducing modification of speciation during steps of sampling and storage by directly trapping in-situ As III and As V. Thus, we worked with commercially available DGT with a resin constituted by ferrihydrite (DGT-FH). We first determined diffusive coefficients of As III and As V by realizing kinetic experiments. Then, we tested several extractants (HNO3 0.8M, HNO3 1M, HCl 4M with acid acetic 10% and hydroxylamine 0.4M) to extract As III and As V with a good reproducibly, by minimizing transformation of As III into As V and by optimizing limit of quantification during analysis by HPLC-ICPMS. To finish we deployed triplicates of DGT-FH during 1 week in summer and in winter on the Ardières and Morcille Rivers to assess inorganic As speciation. Kinetic experiments allowed determining diffusive coefficients for As III (6.52 10-6 cm² s-1) and As V (4.76 10-6 cm² s-1) which are in the same range than those found in literature. Experiments realized in situ showed that DGT-FH was well adapted to trap in-situ As III and As V during the summer campaign. Results highlighted that during this low-water level, As III concentrations in surface waters were ~1 µg/L accounting for 5 to 10 % of total dissolved concentrations. Results of the winter campaign are in progress. This study provided the first attempt to develop a method allowing to extract inorganic species of As previously trapped in-situ by DGT-FH and to analyze by HPLC-ICP-MS. This approach allows us to better characterize, by passive sampling method, temporal exposition of aquatic organisms to As species in surface water.