A 70 day pot experiment was conducted for the cleaning-up of a PCBs-contaminated soil (104 mg kg(-1) soil DW) using bioaugmentation with Burkholderia xenovorans LB400 (LB400) assisted or not by the use of tall fescue (Festuca arundinacea). The total cultivable bacteria of the soil were higher with the presence of plants. Real-time PCR showed that LB400 (targeting 16S-23S rRNA ITS) survived with abundance related to total bacteria (targeting 16S rRNA) being higher with fescue (up to a factor of three). Bioaugmentation had a positive effect on fescue biomass and more specifically on roots (by a factor of three). PCB dissipation (sum of congeners 28, 52, 101, 118, 153, 180) averaged 13 % (bioaugmented-planted) up to 32 % (non bioaugmented-planted), without any significant difference between treatments. Basically our results demonstrated that indigenous bacteria were able to dissipate PCBs (26.2 % dissipation). PCB dissipation was not related to the abundance of LB400 or to the total bacterial counts. Bioaugmentation or fescue altered the structure of the bacterial community of the soil, not the combination of both. Principal component analysis showed that bioaugmentation tended to improve the control of the process (lower variability in PCB dissipation). Opposite to that bioaugmentation increased the variability of the structure of the bacterial community.