Personal exposure to solvents was studied among hairdressers in 28 salons in two regions during two seasons in The Netherlands. Ethanol was used as a marker for solvent exposure. Auxiliary data, such as salon and work characteristics, meteorological conditions and information on the presence of control measures, were collected during the measurements. The average exposure to ethanol was almost a factor of 200 below the occupational exposure limit, but differences in average ethanol concentrations up to a factor of 30 were present between salons. Exposure concentrations were significantly higher on Fridays than on other days of the week. Contrary to expectation, exposures were somewhat lower in the spring than in the summer and in an urban than a semi-rural area. An empirical statistical model based on exposure data collected during the first measurement period appeared not to be valid for the encountered circumstances in the second measurement period. An alternative classification scheme based on two easily obtainable salon and task characteristics was elaborated. This scheme will be applied in an ongoing epidemiological study on reproductive disorders among hairdressers and their offspring.