Noroviruses (NVs) are a predominant cause of viral gastroenteritis outbreaks, but they are difficult to identify because they cannot be cultivated in cell culture. Therefore, reverse-transcriptase polymerase chain reaction (RT-PCR) assays are widely used in the testing of clinical stool specimens for NV. However, testing of perianal swabs in the context of an outbreak is considered to be an insensitive method for identification of NV using the RT-PCR technique. An outbreak of acute gastroenteritis involving 159 soldiers on a training base of the Israel Defense Force in December 1999 allowed us to evaluate this identification method. An epidemiologic investigation, a sanitation survey and a case-control study of exposure to different food items served up to 48 h preceding the outbreak were conducted. Stool samples in the form of post-defecation perianal swabs were collected from 24 ill personnel and three non-ill food handlers. Swabs were tested for the presence of NV by RT-PCR assay. Epidemiologic data were consistent with a pointsource food-borne outbreak which was associated with consumption of fresh vegetable salad in the base mess hall (OR = 4.38, 95% CI 1.51-13.35). Both epidemiologic and clinical features were suggestive of gastroenteritis caused by NV, and upon laboratory analysis perianal swabs from four of 24 cases were positive for NV. The combination of practical specimen collection with perianal post-defecation swabs, and the utilization of a molecular epidemiology approach, can simplify the rapid identification of outbreaks caused by NVs.