Denaturing gradient gel electrophoresis (DGGE) analysis of polymerase chain reaction-amplified genes coding for 16S rRNA was used to assess differences in bacterial community structure as a function of spatial location along the height of two biofilters used to treat a model waste gas stream containing methyl ethyl ketone (MEK). One of the laboratory-scale biofilters was operated as a conventional continuous-flow biofilter (CFB) and the other was operated as a sequencing batch biofilter (SBB). Both biofilters, inoculated with an identical starting culture and operated over a period lasting more than 300 days, received the same influent MEK concentration and same mass of MEK on a daily basis. The systems differed, however, in terms of the fraction of time during which contaminated air was supplied and the overall operating strategy employed. DGGE analysis indicated that microbial community structures differed as a function of height in each of the biofilters. The DGGE banding patterns also differed between the two biofilters, suggesting that operating strategies imposed on the biofilters imparted a sufficiently large selective pressure to influence microbial community structures. This may explain, in part, the superior performance of the SBB over the CFB during model transient loading conditions, and it may open new possibilities for purposely manipulating the microbial populations in biofilters treating gas-phase contaminants in a manner that leads to more favorable treatment performance.