Forty-three cases of serologically confirmed hepatitis A occurred among individuals who ate at restaurant A in Ohio in 1998. Serum samples from all restaurant A employees who worked during the exposure period were negative for IgM antibodies to hepatitis A virus (HAV). A matched case-control study determined that foods containing green onions, which were eaten by 38 (95%) of 40 case patients compared with 30 (50%) of 60 control subjects, were associated with illness (matched odds ratio, 12.7; 95% confidence interval, 2.6-60.8). Genetic sequences of viral isolates from 14 case patients were identical to each other and to those of viral isolates from 3 patients with cases of hepatitis A acquired in Mexico. Although the implicated green onions, which could have come from one of 2 Mexican farms or from a Californian farm, were widely distributed, no additional green onion-associated cases were detected. More sensitive methods are needed to detect foodborne hepatitis A. A better understanding of how HAV might contaminate raw produce would aid in developing prevention strategies.