OBJECTIVE To determine the source of an outbreak of Salmonella javiana and Salmonella oranienburg infections. METHODS Laboratory-based statewide surveillance for Salmonella infections and two separate case-control studies. METHODS Community- and industry-based studies conducted from May through October 1989. METHODS Thirty-one culture-confirmed outbreak-associated cases of S javiana infection and 60 community controls matched for telephone prefix, gender, and age in case-control study I; 50 cases, 100 community controls, and 64 family member controls in case-control study II. RESULTS One hundred thirty-six culture-confirmed cases of S javiana infection and 11 cases of S oranienburg infection were associated with the outbreak in Minnesota. Outbreak-associated cases were also identified in Wisconsin (15 cases), and in Michigan and New York (one case each). Cases were more likely than controls to have consumed mozzarella cheese manufactured at a single cheese plant (plant X) or cheese that had been shredded at processing plants that also shredded cheese manufactured at plant X (odds ratio [OR], 7.2; 95% confidence interval [CI], 1.7 to 23.2; P < .01). The outbreak-associated strains of both serovars were isolated from two unopened 16-oz (0.45-kg) blocks of mozzarella cheese produced at plant X. The most probable numbers of Salmonella organisms in these samples were 0.36/100 g and 4.3/100 g. CONCLUSIONS The potential for bacterial pathogen contamination of cheese during manufacture and processing has important epidemiologic implications, particularly because cheese consumption has recently increased in the United States. Low-level contamination of a nationally distributed food product can cause geographically dispersed foodborne outbreaks that may be difficult to detect.