Root surface caries is an increasing dental health problem. The overall aim of this thesis was to develop an in situ remineralisation model of sub-surface root caries that may be principally used for two purposes: mechanistic studies of the root caries process and the testing of possible caries-preventative agents. The development of this model involved both in vitro and in situ investigations. A necessary prerequisite for the development of such a model was the need to produce sub-surface lesions with histological characteristics similar to that observed in vivo. Following in vitro investigation an acetic acid buffer was chosen as the preferred demineralising system for the production of lesions with well-formed surface layers and uniform areas of sub-surface demineralisation. In addition, fluoride-rich cementum was planed from all hard tissue substrates prior to lesion formation and inclusion in subsequent in vitro and in situ studies. It is an ethical requirement that dental hard tissues be free of pathogens prior to their introduction to the oral environment. From further in vitro studies it was concluded that the disinfection of sound and demineralised dentine for use in in situ caries models should be derived from both the short-term storage of samples in an anti-microbial solution, such as a low concentration thymol-distilled water solution, and through the gamma irradiation of samples at a level that exceeds 4,000 Gy. The chemical processes of artificial sub-surface dentinal lesion remineralisation were investigated in vitro. The presence of low concentrations of ambient fluoride significantly increased the uptake of mineral ions. Although calcium was the preferential mineral ion taken up by the lesions, remineralisation was not stoichiometric. Furthermore, the mineral content and distribution of the lesions at baseline primarily controlled the location and degree of mineral ion uptake. The potential of an 'attached specimen' model of root caries for the remineralisation of sub-surface dentinal lesions was examined in situ. The results of this study lead to the suggestion that larger lesions, which possess highly mineralised surface layers, are less likely to gain mineral intra-orally compared with smaller lesions containing a less mineralised surface zone. The reactive surface area of residual crystals within the lesions surface dictated whether mineral was deposited within the body of the lesions or whether mineral accumulation was confined to the surface layer of the samples. It was concluded that the 'attached specimen' model was suitable for the study of root caries remineralisation. The fluoride content profiles of the sub-surface lesions before and after retention in situ were examined in vitro. The lesions readily acquired fluoride from oral fluids in situ. The highest quantities of fluoride were located within the surface layer of the samples. This was accompanied by a significant increase in the fluoride content of the lesion bodies that may have arisen from the redistribution of fluoride from the surface layer following periods of acid challenge. The biological responses of vital root hard tissues cannot be measured by any of the studies detailed in this thesis. Therefore, in additon to in vitro and in situ studies, the advent and development of techniques such as quantitative laser- and light-induced fluorescence, which allow changes in the mineral content and distribution of carious lesions to be monitored in vivo, are crucial for the further elucidation of the mechanisms of root caries initiation, progression and repair.