Nowadays, pesticides, considering their efficiency on animal and plant parasites, represent considerable progress in food resources management and public health improvement. However, due to intensive use, the three environment compartments (water, air, soil) are contaminated with worrying consequences for human being and ecosystem. In order to conserve and restore the condition of aquatic environments, the European Water Framework Directive (WFD) was adopted in 2000 by the European Union. Indeed, monitoring networks - such as Water Agencies - proceed to regular monitoring, performing spot water sampling up to 12 times a year. But, there are still questions about the spatial and temporal representativeness of this type of sampling. Actually, contamination level in rivers can vary abruptly, depending on season and hydrology at the sampling moment. Thus, contamination peaks could be undetected by monitoring networks, generating a partial snapshot and then estimation of water quality. Passive sampling techniques are proposed and developed to obtain better representativeness of the contamination level, providing average concentration over time with integration of peculiar events like contamination peaks. In this study, the Polar Organic Chemical Integrative Sampler (POCIS) with different king of membranes was considered. This tool, widely used for polar pesticides sampling (0 < log Kow < 4) (Alvarez et al. 2004), allows pre-concentration in situ of analytes and presents an integrative capacity over several week. Its scope and its technical limitations are actually well documented (Lissalde et al. 2011). Furthermore, it was proved that this device may integrate contamination peaks relatively prolonged (about 3 days) for moderately polar compounds (log Kow = 2-3) (Mazzella et al. 2008). Concerning shorter contamination peaks (few hours-few days) generated by punctual pollutions, interrogations about its reactivity (e.g. lag time) is still remaining. This study aimed to show the POCIS reactivity for three different contamination peak patterns, with variable intensity and duration, over a total exposure period of 14 days. For this experimentation under laboratory controlled conditions, fifteen pesticides covering different polarities (log Kow=0-5) and groups (herbicides, fungicides, insecticides) were selected. The data showed different accumulation kinetic profiles, depending on physico-chemical characteristics of the substances and the exposition duration. Some compounds presented a lag effect, or on the contrary a burst effect. For seven compounds, POCIS sampled correctly contamination peaks, which show the ability of this device to sample fleeting events.< Réduire