Abstract To study the physical properties of a bentonite buffer, long-term test (LOT) of buffer material were performed in granitic rock 450 m underground at the Äspö Hard Rock Laboratory (HRL), Oskarshamn, Sweden. The LOT was set up under conditions similar to those in a high-level radioactive waste (HLW) repository, except for the absence of radioactivity and difference in scale, and provided the opportunity of exposing strains of bacteria to conditions realistic to a repository buffer. The main focus was on sulphide-producing bacteria and their ability to survive. Bacteria were chosen for different relevant characteristics. Sulphate-reducing bacteria (SRB) included Desulfovibrio aespoeensis isolated from deep Äspö HRL groundwater, the moderately halophilic bacterium Desulfovibrio salexigens and the thermophilic, spore-forming Desulfotomaculum nigrificans. Aerobic bacteria included Deinococcus radiophilus, a bacterium that can tolerate high doses of radiation and severe desiccation, the chemoheterotrophic bacterium Pseudomonas aeruginosa that frequently occurs in soil, the chemo-organotrophic and chemolithotrophic (hydrogen-utilizing) organism Ralstonia eutropha, the chemoheterotrophic, spore-forming bacterium Bacillus subtilis and the thermophilic spore-forming bacterium Bacillus stearothermophilus. Suspensions of the SRB (anaerobic) and aerobic bacteria were mixed with bentonite clay to give a solution of approximately 100 million bacteria per gram dry weight (gdw −1) clay. The clay with bacteria were subsequently formed into cylindrical plugs with 20 mm length and diameter, and installed in bentonite blocks exposed to low (20–30°C) and high (50–70°C) temperatures. The blocks were installed in the LOT boreholes immediately after incorporation of the bacteria plugs. The experiment was terminated after 15 months. The major outcome was elimination below the detection limits for all except the spore-forming bacteria. All of the three spore formers survived at the low temperature. The numbers remaining were, however, much lower than the ca. 100 million spore-forming bacteria gdw −1 clay initially introduced, so that only between one-100th and one-10 000th of the original number were left. The cell (spore) death rate could therefore be interpreted as being higher than the growth rate, which may have been zero, or close to zero. At the high temperature, the spore-forming SRB D. nigrificans and some of the introduced B. subtilis were the only surviving bacteria. They most probably survived as spores, which are metabolically inactive and do not produce sulphide. A slow but significant death rate of viable cells and spores would eventually lead to the complete eradication of life in the buffer. It has not yet been clarified, however, whether this would occur within the lifetime of a HLW repository.