Fluoride-releasing dental materials are generally believed to reduce or prevent secondary caries. The evidence for this is largely anecdotal and centers on clinical experience with silicate cements and, more recently, with glass-ionomer cements. Unfortunately, corroborating evidence from controlled clinical trials is inadequate to establish precisely how effective these materials are or under what conditions they might be effective. Even less is known about the clinical effectiveness of newer materials that, often, release less fluoride. In vitro model systems have been used to study the effects of dental materials on de/remineralization of surrounding tooth structure. Fluoride-releasing materials have been shown in these models to reduce demineralization of both enamel and dentin compared with a material that does not release fluoride. This is useful from a mechanistic standpoint, but without clinical "standards" to guide model results, it is not possible to define an acceptable level of fluoride release or the length of time such release is required. A limited number of in situ model studies has been conducted, and some fluoride dose-response information has been obtained. These models are closer to the real situation and perhaps provide the best means to define required levels of fluoride release from materials in the absence of adequate clinical information.