The impact of microbial system variability on the biodegradation and transport behavior of a model solute, salicylate, was investigated with a series of miscible displacement experiments. Four systems of increasing complexity were employed: a sterilized, well-sorted sand inoculated with a single bacterial isolate, a sterilized soil inoculated with the same isolate, and two soils, each of which contained an indigenous multiple-population community of bacteria. The experiments were conducted in replicate (three or four experiments per set) and with paired controls. The biodegradation and transport behavior of salicylate exhibited a small degree of variability among the replicates for the two inoculated systems and a relatively large degree of variability for the two indigenous systems. The greater variability observed for the two indigenous systems is attributed primarily to greater variability of microbial system properties, such as initial cell density, metabolic status, and community composition. Values for maximum specific growth rate coefficient, mean lag time, and lag time variance were determined by model calibration to the measured breakthrough curves and compared to values obtained from batch experiments. Reasonable correspondence was observed between the two sets of values for both the inoculated and indigenous systems. The maximum specific growth rate coefficient exhibited a relatively small degree of uncertainty for all four systems, whereas greater uncertainty was associated with the lag time mean and variance. The variability in calibrated parameters among each set of replicate experiments was significantly greater than the uncertainty associated with the individual experiment calibrations and the measured input parameters. These results illustrate that variability inherent to natural microbial systems can cause variability in transport behavior even under controlled laboratory conditions and concomitantly enhance the uncertainty of biokinetic parameters obtained from laboratory studies.