Cell numbers of perchlorate (PRM)- and chlorate (CRM)-reducing microorganisms and the persistence of perchlorate were determined in samples of soils, natural waters, and wastewater incubated under laboratory conditions. Complete perchlorate reduction in raw wastewater and creek water was achieved in 4 to 7 days and 8 to 29 days, respectively, depending on the individual growth substrate (acetate, lactate, citric acid, or molasses) employed. Perchlorate persisted in most mixed cultures developed with 2 g of “pristine” soil, but declined in mixed cultures developed with 100 g of soil. Less than seven days were required to completely reduce perchlorate in cultures started with 10 g of a perchlorate-contaminated soil obtained from a site in Texas. The concentration of PRM was estimated using a 5-tube most probable number (MPN) procedure. To account for discrepancies due to differences in the total number of bacteria (per mass of sample) in the samples, difficulty in removing bacteria from soil samples, and the lack of an unequivocal method to measure total viable cells in these different systems, we normalized our MPN results on the basis of 106 or 109 total bacteria counted using acridine orange direct counts (AODC). There were more PRM in wastewater samples on a per-cell basis (15 to 350 PRM/106-AODC) than in water samples (0.02 to 0.4 PRM/106-AODC). There were also more PRM in soils from sites exhibiting direct evidence of perchlorate contamination (100 to 200 PRM/109-AODC) than from other sites (nondetectable to 0.77 PRM/109-AODC). These results demonstrate that perchlorate-reducing bacteria are present at perchlorate-contaminated sites, and that perchlorate can be degraded by these microorganisms through the addition of different electron donors, such as acetate and lactate.