Abstract The aim was to study the effect of flossing on proximal caries in children grouped according to different combinations of dietary and oral hygiene habits. 140 12–13-year-old children flossed the right or the left lateral region of the mouth once every school-day for 2 years by pulling a waxed floss once up and down through the proximal contacts. Proximal caries was recorded on bitewing films, taken at the baseline and at the 2-year registration. Diet history was obtained four times during the period. Oral hygiene was classified according to the number of bleeding points registered with a feather-loaded probe at the final examination. Caries increments on the control side during the experimental period were calculated for groups with the same habits. A logical border with respect to the amount of new caries was then established between suitable and unsuitable dietary habits as well as between sufficient and insufficient oral hygiene. No statistically significant difference was found between the control and experimental subgroups in the different combinations of dietary and oral hygiene habits, neither with respect to the whole material nor comparable contralateral surfaces. The numerical differences were so small that the technique of flossing studied here cannot be recommended.