A case-control study was carried out to investigate an outbreak of acute gastroenteritis among a military detachment stationed in a rural area of Castellón, España. The purpose of the study was to determine the causes of the outbreak and develop control measures. Of the 153 men in the detachment, 135 were included in the study. Between 9 and 11 August 1993, 45 cases were reported; the patients' average age was 19.2 +/- 1.5 years. The attack rate was 33.3%. The clinical picture was dominated by the following symptoms: diarrhea (76%), vomiting (67%), nausea (67%), and abdominal pain (28%). The median duration of symptoms was one day, and that of the incubation period was 33 hours. Only one patient required hospitalization and all of them recovered. Salmonella richmond (6.7: and :1.2) was isolated in 5 of the 14 stool cultures performed. An association was also discovered between the illness and consumption of water from an aqueduct that flowed near the camp. A logistic regression model showed that consumption of water from this source remained associated with cases after adjusting for age and the consumption of various foods (odds ratio = 96.5; 95% confidence interval, 11.4-814.4). The risk of suffering from the illness rose with the amount of water consumed (chi 2 trend test = 65.4, P < 0.0001). Chemical and bacteriological analyses of the aqueduct water indicated the presence of fecal contamination. The aqueduct had not been subject to sanitary monitoring, even though the water was used to irrigate agricultural crops. The widespread presence in the environment of species of Salmonella was demonstrated. Health education and microbiological studies of water courses can be of great value in preventing such epidemics.