To study the efficacy of sodium fluoride varnishes and a NaF solution in remineralization of enamel, 120 slabs of non-carious human enamel enamel were presoftened for 6 h and randomly divided into six groups. The slabs were stored in synthetic saliva for 9 days, except for a daily 30-min immersion in 0.1 M lactic acid-NaOH buffer. During the 9-day period, one group of the slabs received no treatment, and the rest were treated once or three times with 2.3% or 1.1% sodium fluoride varnish Duraphat, or nine times with a 0.1% NaF solution. Finally, the slabs were demineralized for 1 h, and the amount of dissolved Ca and F was determined. Microhardness of enamel was determined initially, after presoftening, after the 9-day period, and after the 1-h demineralization. All fluoride treatments prevented enamel softening almost completely during the 9 days, but the control slabs softened markedly. Fluoride varnishes were more effective than NaF solution. Three applications of 2.3% Duraphat were slightly more effective than any of the other varnish treatments, but one treatment with 2.3% varnish was not more effective than treatments with 1.1% varnish. Enamel treated three times with 1.1% varnish showed the greatest acid resistance during the 1-h demineralization. The results suggest that the efficacy of the varnish was not proportional to the fluoride concentration but rather to the number of applications. Fluoride uptake by enamel was greatest with the most concentrated varnish. Enamel solubility was not, however, directly proportional to the fluoride content of enamel.