A previously described bioassay using a mer-lux gene fusion for detection of bioavailable mercury was applied for the estimation of the bioavailable fraction of mercury in soil. The bioavailable fraction is defined here as being part of the water leachable fraction. Due to masking of light emission of soil particles leachates had to be cleaned prior to assays. Filtration of leachates through nitro-cellulose filters using pressure resulted in an underestimation of bioavailable mercury. Gravity filtration and centrifugation showed elevated (as compared with untreated leachate) and very similar responses. The utility of the mer-lux biosensor assay was tested by relating measurements of bioavailable and total mercury to the response of the soil microbial community to mercury exposure. Two different soil types (an agricultural and a beech forest soil) were spiked with 2.5 µg Hg(II) g^-1 in microcosms and the frequency of mercury resistant heterotrophs and changes in community diversity, defined as the number of different 16S rDNA bands observed in DGGE gels, were monitored. In the agricultural soil the initial concentration of bioavailable mercury was estimated to be 40 ng g^-1. This concentration did not change during the first 3 d and coincided with increased degrees of resistance and a decrease in diversity. The concentration of bioavailable mercury decreased subsequently rapidly and remained just above the detection level (0.2 ng g^-1) for the remainder of the experiment. As a possible consequence of the decreased selection pressure of mercury, the resistance and diversity gradually returned to pre-exposure amounts. In the beech forest soil the concentration of bioavailable mercury was found to be about 20 ng g^-1 throughout the experiment. This concentration did not at any time result in changes in resistance or diversity. This study showed that the bioassay using the mer-lux biosensor is a useful and sensitive tool for estimation of bioavailable mercury in soil.