Two ozone (10028156) monitoring procedures were compared: the alkaline potassium-iodide (AKI) impinger method; and a direct reading chemiluminescence monitor, the Analytical Instrument Development (AID) Incorporated, Model 560 ozone analyzer. The accuracy and precision of each method were first tested in the laboratory using an ozone generator calibrated by ultraviolet photometry. Field sampling was then conducted during metal inert gas welding of mild steel, under different ventilation and welding amperage conditions. Air samples were taken through side by side closed face filter cassettes mounted on the welder's helmet. All sampling was performed simultaneously for the two techniques. In the laboratory tests at ozone concentrations of 0.052, 0.201, 0.802, and 1.61 parts per million (ppm), the AID monitor showed good accuracy and precision while more variability was observed with the AKI impinger. In the field tests with a high ventilation rate and low amperage, the AID monitor and AKI impinger recorded ozone levels of 0.072 to 0.090 and 0.042 to 0.048ppm, respectively. The mean agreement between the methods was 56.7 percent. At reduced ventilation rates and increased amperage the ozone levels were extremely high. Concentrations in this case registered with the AID monitor ranged from 1.32 to 4.80ppm, while the AKI impinger indicated much lower levels of 0.131 to 0.541, and the mean agreement between the two methods was only 11.9 percent. The authors conclude that negative interference created by contaminants may be the cause of the low readings obtained with the AKI impinger method during the field tests. They recommend that the validity of use of the impinger method in industrial situations should be reevaluated, and emphasize the merit of the AID monitor as a field instrument.