The expiratory bradypnoea indicative of upper airway irritation in mice was evaluated during a 60-min oronasal exposure to increasing concentrations of chlorine and nitrogen trichloride. The airborne concentration resulting in a 50% decrease in the respiratory rate of mice (RD50) was calculated for each chemical. Chlorine and nitrogen trichloride showed dissimilar concentration-response curves. While the maximal response of nitrogen trichloride was reached in 10 min, the maximal response of chlorine was reached between 45 and 60 min of exposure. The results showed both chemicals to have an irritant potency of the same order of magnitude. The RD50 values of chlorine and nitrogen trichloride were 3.5 and 2.5 ppm, respectively. On the basis of a TLV-STEL (threshold limit value for short-term exposure limit) equal to 0.1 RD50 and a TLV-TWA (time-weighted average) equal to 0.03 RD50, the current TLVs for chlorine seem too high (1 and 0.5 ppm, respectively) and should be reduced to 0.5 and 0.1 ppm, respectively. For nitrogen trichloride, 0.3 ppm and 0.1 ppm are proposed as TLV-STEL and TLV-TWA, respectively.