A pilot study to test methods of estimating personal exposures to toxic substances and corresponding body burdens was carried out between July and December 1980. Individual exposures to about a dozen volatile organic compounds in air and drinking water were measured for nine volunteers in Bayonne and Elizabeth, New Jersey, and for three volunteers in Research Triangle Park, North Carolina during three 3-day visits over the 6-month period. Breath samples were also collected from all subjects on each visit. Composite food samples were collected in each locality. Sampling and analytical methods for air, water, food, and breath were evaluated and found generally capable of detecting concentrations as low as 1 microgram/m3 in air and breath, and 1 ng/g in water and food. About 230 personal air samples, 170 drinking water samples, 66 breath samples, and 4 food samples (16 composites) were analyzed for the target chemicals. Ten compounds were present in air and eight were transmitted mainly through that medium. The two target trihalomethanes (chloroform and bromodichloromethane) were predominantly transmitted through water and beverages. Food appeared to be a minor route of exposure, except possibly for trichloroethylene in margarine. Seven compounds were present in more than half of the breath samples. Diurnal and seasonal variations were noted in air and water concentrations of some compounds, with summer levels generally higher. For some chemicals, weekday air exposures were significantly higher than weekend exposures. Some, but not all, of the potentially occupationally exposed individuals had significantly higher workplace exposures to several chemicals. Distributions of air exposures were closer to log normal than normal for most chemicals. Several chemicals were highly correlated with each other in personal air samples, indicating possible common sources of exposure.