The dynamics of a multispecies biofilm population in a laboratory-scale trickle-bed bioreactor for the treatment of waste gas was examined. The model pollutant was a VOC-mixture of polyalkylated benzenes called Solvesso 100. Fluorescence in-situ hybridization (FISH) was applied in order to characterise the population composition. The bioreactor was operated under transient conditions by applying pollutant concentration shifts and a starvation phase. Only about 10% of the biofilm mass were cells, the rest consisted of extracellular polymeric substances (EPS). The average fraction of Solvesso 100-degrading cells during pollutant supply periods was less than 10%. About 60% of the cells were saprophytes and about 30% were inactive cells. During pollutant concentration shift experiments, the bioreactor performance adapted within a few hours. The biofilm population exhibited a dependency upon the direction of the shifts. The population reacted within days after a shift-down and within weeks after a shift-up. The pollutant-degraders reacted significantly faster compared to the other cells. During the long-term starvation phase, a shift of the population composition took place. However, this change of composition as well as the degree of metabolic activity was completely reversible. A direct correlation between the biodegradation rate of the bioreactor and the number of pollutant-degrading cells present in the biofilm could not be obtained due to insufficient experimental evidence.