In a municipality near Copenhagen, Denmark, where fortnightly fluoride rinses with 0.2% neutral sodium fluoride had been performed for more than a decade, 1306 children from kindergarten through sixth grade were stratified by school and grade and randomly distributed into two groups. One group continued with the fluoride rinses, the other group had the fluoride solution replaced with distilled water. Both solutions were slightly flavored. One thousand and eighty-three children completed the three-year trial. Caries was recorded clinically by the dentists in the municipal dental service using criteria common for the Child Dental Health Services, and on bitewing radiographs by one of the authors applying criteria developed by Gröndahl et al. Permanent molars and premolars were included in the study. Clinically, caries increment in the two groups was the same, with 94% of caries lesions and fillings situated in pits and fissures. According to the radiographs, caries progression in the water group was higher than in the fluoride group. This difference was statistically significant for the surfaces erupting during the study (p < 0.05).