The use of fluoride-releasing restoratives such as glass ionomer cements (GICs) has increased during the last decade. The antibacterial effect of released fluoride is thought to be a possible caries-preventive effect of these restorations. In this study fluoride concentrations in plaque on 1-year old resin-modified GIC, compomer and resin composite restorations were compared intraindividually and related to the occurrence of caries-associated bacteria. Plaque from class III restorations of the three restorative materials and from a proximal enamel surface in 18 individuals was analysed. Low fluoride levels were detected in all the samples, while the resin-modified GIC samples showed significantly higher amounts. The distribution of oral streptococci, mutans streptococci and lactobacilli did not differ significantly among the surfaces and did not correlate to the fluoride levels in the samples. A good correlation was found between the counts of mutans streptococci in saliva and their proportions in the plaque. The results indicate that the fluoride concentrations released in vivo from 1-year-old restoratives are not high enough to affect the plaque levels of the caries-associated bacteria mutans streptococci and lactobacilli.