A laboratory‐scale peat biofilter was used for ethyl acetate removal from an air stream over a period lasting 1 yr. In a first stage, the biofilter was operated under continuous mode: a maximum elimination capacity of 400 g m−3 h−1 was obtained, and ethyl acetate was efficiently degraded with empty bed residence time (EBRT) as short as 22 s. The estimated yield coefficient, determined from the carbon dioxide production, resulted in 0.42 g dry biomass produced per gram of ethyl acetate consumed. The living and the dead cell concentrations were also monitored. The dead cell percentages varied between 18 and 85%, progressive increases in the dead cell percentages were achieved as EBRT decreased and inlet load increased. In a second stage, operation under intermittent loading (16 h/day or 2‐h‐on/2‐h‐off cycles) was performed to investigate the effect of industrial practices such as overnight and weekend shutdowns, intermittent emission or inlet concentration and flow rate peaks. System performance was not significantly affected by changes, demonstrating the process viability for industrial applications. The intermittent loading of 16 h/day resulted in lower penetration of contaminant in the bed in comparison with continuous loading, but the 2‐h‐on/2‐h‐off caused a greater pollutant breakthrough detected at the first quarter of the biofilter, with a high increase in the dead cell density. Respirometric monitoring showed that biomass was fully activated in 1–2 h after night or weekend starvations. Finally, a 15 days starvation period was applied and afterwards operation was restarted. The system performance was fully restored without noticeable reacclimation period and dead cell density recovered normal values. © 2007 American Institute of Chemical Engineers Environ Prog, 2007