The evolution of impure rocksalt samples under irradiation has been investigated and compared to the evolution of pure halite, concerning the radiolytic release of gases. Four different types of rocksalt (pure halite, anhydrite-bearing, sylvinite-bearing, and marneous) have been irradiated with spent fuel at doses ranging from 10 4 Gy to 10 7 Gy, under synthetic air or helium at 50'C, 150'C, and 200'C. Marneous and anhydrite-bearing salt produced higher amounts of gas than the two other types, especially for H 2 , CO, and CH 4 . Radiolytic gas production essentially originates from the decomposition of traces of different kind of impurities : fluid inclusions, which can be composed of brine and/or organic fluids and gases; minerals such as carbonates, sulfates, or hydrated minerals; organic matter, in general (kerogen). Pure halite decomposes only at very high doses (more than 10 7 Gy) giving off some Cl 2 and other corrosive gases. Gas production increases steadily with dose for H 2 , CO, and CH 4 , contrary to what is observed for the pure Asse salt, where H 2 and CO, respectively, peak at 106, and 105 Gy. The increase, with dose, in the production of CH 4 , C 2 H 6 , and C 3 H 8 is far higher for marneous salt and higher for anhydrite-bearing salt. The decrease of the 0 2 /N 2 (oxygen consumption) ratio in the irradiation atmosphere is in the following order: marneous>anhydrite>halite=sylvinite. The overall effect of 'y-irradiation on impure rocksalt can be described as follows : destruction of organic matter that produces chemically reduced gases (H 2 , CO, hydrocarbons), after an initial stage of oxygen consumption, which is reached more or less quickly, depending on the availability and quantity of organic matter. The use of impure, i.e, organic matter-bearing salt, as host-rock or engineered barrier for nuclear wastes, may be questioned, because of the high amount of gases produced in the immediate vicinity of the container (explosion hazards, pressure build-up). The present study will be followed by a modelling of the in-situ radiolytic gas production, taking into account: 1) the dose distribution with time, 2) the temperature distribution with time, and 3) the geometry of the repository, and will allow us to estimate the potentiel safety impact of radiolytic gas production in the case of a repository emplaced in impure rocksalt.