A comparison was made between the demineralization of enamel and dentine with and without abraded surfaces. This was done in a pH-cycling experiment for different demineralization/remineralization ratios - in the range from 1:1 to 1:4 - and for different fluoride additions (up to 2 ppm) in solution. A new automatic pH-cycling system, in which the de- and remineralization solutions have a constant composition during the de- and remineralization cycles, was used to create mineral loss in human dentine and enamel. Changes in mineral content were monitored by means of longitudinal microradiography. A linear correlation was found between the amount of mineral lost and the total demineralization time for both dentine and enamel. The demineralization rates were comparable for abraded enamel and dentine and for polished enamel and dentine, and this rate was roughly doubled by the removal of the outer surface for both tissues. This showed that the presence of the outer surface is equally important to dentine and enamel. Under the pH-cycling conditions used, a logarithmic relation was found between the addition of fluoride and the decrease in demineralization for both enamel and dentine. The inhibitory effect of fluoride on demineralization was most pronounced on abraded enamel, followed by pumice-polished enamel, abraded dentine and pumice-polished dentine. About 2 ppm fluoride was needed under the conditions used to stop enamel demineralization completely; in the case of dentine, however, this amount of added fluoride did not inhibit demineralization.