Pseudomonas fluorescens strains are used in agriculture as plant growth-promoting rhizobacteria (PGPR). Nontarget effects of released organisms should be analyzed prior to their large-scale use, and methods should be available to sensitively detect possible changes in the environments the organism is released to. According to ecological theory, microbial communities with a greater diversity should be less susceptible to disturbance by invading organisms. Based on this principle, we laid out a pot experiment with field-derived soils different in their microbial biomass and activity due to long-term management on similar parent geological material (loess). We investigated the survival of P. fluorescens CHA0 that carried a resistance toward rifampicin and the duration of potential changes of the soil microflora caused by the inoculation with the bacterium at the sowing date of spring wheat. Soil microbial biomass (C(mic), N(mic)) basal soil respiration (BR), qCO(2), dehydrogenase activity (DHA), bacterial plate counts, mycorrhiza root colonization, and community level substrate utilization were analyzed after 18 and 60 days. At the initial stage, soils were clearly different with respect to most of the parameters measured, and a time-dependent effect between the first and the second set point were attributable to wheat growth and the influence of roots. The effect of the inoculum was small and merely transient, though significant long-term changes were found in soils with a relatively low level of microbial biomass. Community level substrate utilization as an indicator of changes in microbial community structure was mainly changed by the growth of wheat, while other experimental factors were negligible. The sensitivity of the applied methods to distinguish the experimental soils was in decreasing order N(mic), DHA, C(mic), and qCO(2). Besides the selective enumeration of P. fluorescens CHA0 rif(+), which was only found in amended soils, methods to distinguish the inoculum effect were DHA, C(mic), and the ratio of C(mic) to N(mic). The sampling time was most sensitively indicated by N(mic), DHA, C(mic), and qCO(2). Our data support the hypothesis-based on ecosystem theory-that a rich microflora is buffering changes due to invading species. In other words, a soil-derived bacterium was more effective in a relatively poor soil than in soils that are rich in microorganisms.