For a proper analysis of the potentially causal relationship between exposure to volatile organic chemicals (VOCs) in drinking water and health events, it is essential to know T1, the time when exposure started, and C = f(T), which is the change of the VOC concentration C as a function of time T and the total accumulated exposure (TAE) to VOCs to which an individual was exposed. In the typical situation of incidentally detected pollution of groundwater, no such information is available. This paper describes the development of a method for estimating T1, C = f(T), and TAE as part of an epidemiologic study of the health effects of VOC contamination of an aquifer serving public and private wells. Pooled test results of city wells, tested periodically since 1981, provided the data base for developing a statistical model for estimating C = f(T). This model was then applied to private wells, for which the data of only one water sample were available, to retrospectively estimate their T1. The best-fitting model was a multiple linear regression equation consisting of the natural logarithm of the VOC concentration as the response variable, with the time of sampling, the distance of the wells from the source (expressed as coordinates), the well depth, and the well capacity as determinants. The TAE was calculated by integrating the area under the time-concentration curve.