Tobacco smoke-related air pollutant levels were studied in ten Finnish restaurants. Markers of tobacco smoke were measured together with other compounds typical of tobacco smoke and indoor air. The measurements were carried out at stationary sampling points in smoking and non-smoking areas of the restaurants in 2005-2006, when at least half of the service area had to be non-smoking according to the Finnish Tobacco Act. The average concentrations (geometric mean, microg/m3) of the 16 airborne contaminants measured in the smoking area were: nicotine 18.1; toluene 10.6; isoprene 10.2; m,p-xylene 5.0; limonene 4.8; benzene 3.3; furfuryl aldehyde 3.2; 1,3-butadiene 2.7; 3-ethenylpyridine (3-EP) 2.5; phenol 2.1; ethyl benzene 1.7; pyridine 1.6; o-xylene 1.5; 3-picoline 1.4; styrene 1.2; and naphthalene 0.45. A good correlation (r=0.90-0.99, p<0.001) was obtained between tobacco-specific markers (3-EP and nicotine) and 1,3-butadiene, isoprene, pyridine, furfuryl aldehyde, 3-picoline, phenol, and styrene. A poor or no correlation (r=0.19-0.60) was obtained between 3-EP or nicotine and the rest of the compounds. The average concentrations of all compounds were significantly lower in the non-smoking area than in the smoking area (p<0.05). In the non-smoking area, the average concentration of 3-EP was 0.35 microg/m3 and that of nicotine 1.6 microg/m3. In three restaurants, the area design and ventilation were effective: the average level of 3-EP in the non-smoking section was <3% from that in the smoking section. In the other restaurants, tobacco smoke was spreading more freely and the corresponding value was 14-76%. A sensitive method was applied for the measurement of airborne 1,3-butadiene. The air samples were collected into Carbopack X adsorption tubes and analysed by thermal desorption/gas chromatography/mass selective detection. The precision of the method was 4.2% (at 100 ng/sample) and the limit of quantification 0.02 microg/m3.