The aim was to investigate interactions between enamel and dentine at low pH under conditions simulating those at the enamel-dentine junction. Sound enamel blocks were demineralised in acid-gel systems, at pH 4.6, either in isolation, next to one, or in the middle of two, abutting dentine blocks. The gels were initially infinitely undersaturated with respect to enamel. In a second study, enamel blocks containing pre-formed lesions were placed in acid-gel systems, at pH 5.0, either in isolation or next to dentine blocks. The systems were initially either partially or infinitely undersaturated. In the partially saturated systems, calcium and phosphate concentrations were representative of plaque fluid. In the first study, demineralisation of enamel next to one dentine block was reduced in inverse proportion to the distance from the dentine. Demineralisation of enamel between two dentine blocks was retarded markedly across the whole block. In the second study, in the partially saturated systems, enamel lesions next to dentine blocks remineralised, whereas those in isolation demineralised further. We suggest that diffusion of dissolved dentine mineral over the enamel in the infinitely undersaturated system was sufficient to reduce undersaturation, thus retarding demineralisation, and that in the partially saturated systems, dentine dissolution together with the added calcium phosphate caused remineralisation of enamel lesions. Fluoride released from dissolving dentine may have augmented these effects. Different rates of demineralisation in enamel and dentine, or enamel remineralisation with concurrent dentine demineralisation, enabled by differences in their solubilities, could help explain the progression of so-called 'hidden caries'.