Ten healthy subjects were exposed to passive smoking at a high level corresponding to 25-30 ppm CO in the ambient air for 3 h. All subjects were exposed at the same time in a climatic chamber especially designed for exposure experiments. Despite an identical exposure rate considerable interindividual variability of subsequent nicotine and cotinine levels in saliva, plasma and 24-h urine were observed. This variability was more prominent in nicotine than in cotinine levels. The kinetic pattern as reflected by saliva levels for up to 24 h was consistent with previous data found in active smokers. Nicotine levels found in saliva were markedly influenced by repeated sampling. This was not the case for cotinine levels. With regard to laboratory techniques RIA seems to be more sensitive than gas chromatography (GC). The results of this study suggest that measuring cotinine levels in 24-h urine with RIA is presently the most sensitive and reliable criterion for estimating exposure to passive smoking and for validating questionnaires or interviews about short-term exposure to passive smoking.