A total of 601 sera from groups of workers exposed to heavily and moderately contaminated humidifiers respectively were examined by the double diffusion test for precipitating antibodies to humidifier extracts. Clinical information was obtained using a standardized questionnaire. Skin-prick tests to control, to three common inhalant antigens and to an extract from the humidifier were performed in 103 subjects exposed to a heavily contaminated humidifier. In this environment a strong inverse relationship between current smoking and precipitins was detected (P < 0.001). Pack years smoked had no bearing on precipitin level. In the larger study there was a clear dose response effect of smoking on precipitins. The effect of smoking appeared to reverse within 3 years. Duration of exposure also had a major effect on the development of precipitins with a clear ‘dose response’ again being seen in the larger study. These relationships did not apply in the groups exposed to moderately contaminated humidifiers. Levels of airborne antigen were measured using a competitive ELISA inhibition assay and were found to be much lower in sites with moderately contaminated humidifiers than at sites with heavily contaminated humidifiers. Our results suggest that the airborne antigen level is crucial in determining the pattern and strength of precipitin response in smokers and non-smokers exposed to contaminated humidifiers.