An epidemiological assessment of differences in caries and fluorosis prevalences between children in Truro (< 0.1 ppm) and Kentville (fluoridated at 1.1 ppm in 1991), Nova Scotia, Canada, was completed in 1991. Out of a total of 429 children, in grades 5 and 6, in the two towns in 1991, 219 (51%) were examined. Parents answered a self-administered questionnaire investigating the sources of drinking water used by the children since birth, residence history, use of fluoride supplements, dentifrices, and other fluoride products during the first 6 yr of the life. The examination criteria differentiated between non-cavitated and cavitated carious lesions. Dental fluorosis was measured using the TSIF index. Examiner agreement was excellent. Of the children examined, 80 (36.5%) drank water (fluoridated or non-fluoridated) from municipal water systems during the first 6 yr of life. The children were assigned into five groups based upon residence history and exposure to fluoridated water during the first 6 yr of life. The percentage difference in mean DMFS scores between children in the fluoridated and non-fluoridated groups is 17% (delta DMFS1 = 0.7) when non-cavitated carious lesions are included and 39% (delta DMFS2 = 1.1) when they are excluded. The differences are not statistically significant. The significant risk factors associated with the DMFS1 and DMFS2 scores identified by a stepwise multiple regression analysis are: education level of the father, gender, and number of years of reported use of toothpaste during the first 6 yr of life. Dental fluorosis (mainly TSIF score of 1) was present in 41.5% and 69.2% of the children in the non-fluoridated and fluoridated groups, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)