A 21-month trial of the caries-preventive effect of a 0.2% NaF mouthrinse included monitoring changes in the concentration of F in normal enamel. Subsamples of 25 children were randomly selected from groups receiving either normal dental treatment (Control group); additional fortnightly professional prophylaxes (Test group); or prophylaxes plus fortnightly 2-min fluoride rinses (Test + F group). At the start of the trial the sound labial surface of tooth 24 (occasionally 23 or 22) was etched with 2 N HCL and the etchant analysed for F, Ca and phosphate. The contralateral tooth was sampled at the end of the trial. The caries increments (DFS) for the three groups over 21 months were Control, 2.88; Test, 2.51 and Test + F, 1.77, the last being significantly less than the others (P less than 0.05). The average sound enamel surface F concentrations were similar for the three groups at the start and the end of the trial. F parameter calculations showed that all differences were non-significant. Thus this factor could not explain the reduced caries increment due to F mouthrinsing. F biopsies for the purpose of predicting likely clinical effectiveness of neutral 0.2% NaF mouthrinsing will find little application in public health dentistry.