Bacterial activities, estimated using the thymidine and leucine incorporation techniques, were measured in soil artificially contaminated with five different metals (Cd, Cu, Ni, Pb, Zn) at two different concentrations and in an uncontaminated control. The activity of the bacteria was followed during 14 months of incubation under laboratory conditions. After an initial decrease, which was correlated to the level of metals added, the bacterial activity recovered. Peak activities, much higher than those of the uncontaminated control, were detected after different incubation times for the different metal treatments. Cu contaminated soils showed peak activities after 1 month of incubation, while for Zn, Pb, and especially Ni, peak activities were not found until up to after 1 year of incubation. The leucine/thymidine incorporation ratio (mol/mol) did not differ over time, but significant differences were found between different metal treatments. Thus, while the uncontaminated control, and Cd, Zn and Ni treated soils had ratios around 20, the Cu contaminated soils had high ratios, around 50. The Pb contaminated soils, on the other hand, had low ratios (around 7). This was interpreted as an indication of different bacterial communities in the different treatments, where the Cu treated soils differed most from the other soils, while Cd, Zn or Ni application gave rise to similar communities. Other techniques to characterize soil microorganisms have earlier indicated the same effect of the metals on the bacterial community.