Contamination of drinking water with petroleum products is an increasingly common problem. Physicians are often asked to advise patients about such exposures. This study assessed household exposure from gasoline-contaminated drinking water in a New England household. A sampling strategy was designned to estimate inhalation and ingestion exposure to benzene and three other aromatic hydrocarbons typically found in gasoline-contaminated water. The estimated inhaled doses of all agents were similar to the estimated ingested dose. Over half the inhaled dose of all four agents was associated with shower activities as was over half the estimated total dose by all routes of exposure. Under these conditions, discontinuing ingestion of water contaminated with these agents may decrease the dose of benzene by less than one third, whereas discontinuing both ingestion and showering may decrease the dose of benzene by over three quarters. This limited study suggests that routes of exposure other than ingestion are important and should receive attention in the regulatory and risk-assessment process.