Breakthrough times of a commercially-available respirator cartridge packed with a small amount of activated carbon were determined by three different flow patterns: workers' respiratory patterns, conventional steady-state flows, and three standard patterns consisting of rectangle, sine, and triangle waves. A mechanical breathing simulator was used to reproduce the three standard wave patterns as well as the workers' respiratory patterns, which had been recorded from three workers of a paint manufacturing plant wearing a gas mask equipped with a respirator cartridge. The breakthrough test was carried out with 300 ppm carbon tetrachloride vapor at average flow rates of the respiratory and standard patterns equivalent to the steady-state flow rates. The breakthrough times determined by the workers' respiratory patterns were found to be shorter than those by a steady-state flow of the equivalent flow rate, and to be as short as those by sine and triangle wave patterns. It was considered that shorter breakthrough times in the workers' respiratory patterns could be attributed to a faster maximum inspiratory flow rate than that in the case of steady-state flow. The present results suggest that the passing time of CCl4 molecules through the thin activated carbon layer of respirator cartridge may be too short to attain an adsorption equilibrium.