Contamination of groundwater with petroleum hydrocarbons and additives, such as methyl tert-butyl ether (MTBE), is often linked to the leaking product distribution system of gas stations. In very few cases is it know if and when a leak occurred and how much product was released to the environment. In the absence of direct evidence, a careful analysis of the available data, such as contaminant breakthrough at receptor wells or discrepancies in the product inventory data, may provide evidence about the nature of the release, its timing and magnitude. Using a MTBE contamination site in the formerly glaciated New England region as an example, two possible release scenarios (slow, long-term release vs. spill) were examined. Of the two scenarios, the slow release could be ruled out as the sole source even though there was no direct evidence for a spill. The analysis of hydraulic test results together with chemical data further permitted to estimate when such an undocumented spill might have occurred. Analyses of the data also allowed these results to be compared to that of a prior transport and fate modeling study. Good agreement and consistency for contaminant travel times was confirmed. This forensic analysis demonstrates that applying basic hydrogeologic principles can aide in the reconstruction of contamination events while providing more readily understandable and defendable evidence relative to complex models. Conceptually, the approach described herein is transferable to other sites with similar hydrogeologies.