Twenty volatile organic compounds were measured in the personal air and drinking water of 350 New Jersey residents in the fall of 1981. Two consecutive 12-hour integrated personal air samples and two tap water samples were collected from each participant. At the end of the 24-hour monitoring period, each participant supplied a sample of exhaled breath. Simultaneous outdoor samples were collected in 100 residential locations in two cities. Eleven compounds were present much of the time in air, but only four (the trihalomethanes) in water; wide ranges of exposures (three to four orders of magnitude) were noted for most compounds. Ten of 11 compounds displayed significant correlations between air exposures and breath concentrations; the 11th (chloroform) was correlated with drinking water exposures. It was concluded that breath measurements are a feasible, cost-effective, and highly sensitive way to determine environmental and occupational exposures to volatile organic compounds.