In a previous study, we explored the use of acetate, lactate, molasses, Hydrogen Release Compound (HRC, which is based on a biodegradable poly-lactate ester), methanol and ethanol as carbon source and electron donor to promote bacterial sulfate reduction in batch experiments, this with regards to applying an in situ metal precipitation (ISMP) process as a remediation tool to treat heavy metal contaminated groundwater at the site of a nonferrous metal work company. Based on the results of these batch tests, column experiments were conducted with lactate, molasses and HRCI as the next step in our preliminary study for a go-no go decision for dimensioning an on site application of the ISMP process that applies the activity of the endogenous population of sulfate-reducing bacteria (SRB). Special attention was given to the sustainability of the metal precipitation process under circumstances of changing chemical oxygen demand (COD) to [SO4(2-)] ratios or disrupted substrate supply. To optimize the ISMP process, an insight is needed in the composition and activity of the indigenous SRB community, as well as information on the way its composition and activity are affected by process conditions such as the added type of C-source/ electron donor, or the presence of other prokaryotes (e.g. fermenting bacteria, methane producing Archaea, acetogens). Therefore, the biological sulfate reduction process in the column experiments was evaluated by combining classical analytical methods [measuring heavy metal concentration, SO4(2-)-concentration, pH, dissolved organic carbon (DOC)] with molecular methods [denaturing gradient gel electrophoresis (DGGE) fingerprinting and phylogenetic sequence analysis] based on either the 16S rRNA-gene or the dsr (dissimilatory sulfite reductase) gene, the latter being a specific biomarker for SRB. All carbon sources tested promoted SRB activity, which resulted within 8 weeks in a drastic reduction of the sulfate and heavy metal contents in the column effluents. However, unexpected temporal decreases in the efficiency of the ISMP process, accompanied by the release of precipitated metals, were observed for most conditions tested. The most dramatic observation of the failing ISMP process was observed within 12 weeks for the molasses amended column. Subsequent lowering the COD/ SO4(2-) ratio from 1.9 to 0.4 did not alter the outcome of sulfate reduction and metal precipitation efficiency in this set-up. Remarkably, after 6 months of inactivity, bacterial sulfate reduction was recovered in the molasses set up when the original COD/ SO4(2-) ratio of 1.9 was applied again. Intentional disruption of the lactate and HRC supplies resulted in an immediate stagnation of the ISMP processes and in a rapid release of precipitated metals into the column effluents. However, the ISMP process could be restored after substrate amendment. 16S rDNA-based DGGE analysis revealed that the SRB population, in accordance with the results of the previously performed batch experiments, consisted exclusively of members of the genus Desulfosporosinus. The community of Archaea was characterized by sequencing amplicons of archaeal and methanogen-specific PCR reactions. This approach only revealed the presence of non-thermophilic Crenarchaeota, a novel group of organisms which is only distantly related to methane producing Euryarchaeota. DGGE on the dsrB genes was successfully used to link the results of the ISMP process to the community composition of the sulfate reducing bacteria. In the case of an intentional disruption of substrate supply, the ISMP process failed most likely because the growth and activity of the indigenous SRB community stopped due to a lack of a carbon and electron donor. On the other hand, the cause of the sudden temporal shortcomings of the ISMP process in the presence of different substrates was not immediately clear. It was first thought to be the result of competition between methanogenic prokaryotes (MP) and sulfate reducers, since the formation of small amounts of CH4 (0.01-0.03 ppm ml(-1) was detected. However, the results of molecular analyzes indicate that methanogens do not constitute a major fraction of the microbial communities that were enriched in the column experiments. Therefore, we postulate that the SRB population becomes inhibited by the formed metal sulfides. Our results indicate that the ISMP process is highly dependent on SRB-stimulation by substrate amendments and suggest that this remedial approach might not be viable for long-term application unless substrate amendments are continued and environmental conditions are strictly controlled. This will include the removal of affected aquifer material from the metal precipitation zone at the end of the remediation process, or removal of metal precipitates when the microbial activity decreases. Additional tests are necessary to investigate what will happen when clear groundwater passes through the reactive zone while no more C-sources are amended and all indigenous carbon is consumed. Also, the effects of dramatic increases in sulfate- or HM-concentrations on the SRB-community and the concomitant ISMP process need to be studied in more detail.