Aims: To analyze predictors of bruxism, other oral parafunctions, and tooth wear in a group of subjects who had been examined 20 years earlier. Methods: Originally, 402 randomly selected 7-, 11-, and 15-year-old subjects were examined clinically and by means of a questionnaire. Twenty years after the first examination, 94% of the original group could be traced, and 320 (85%) completed and returned the questionnaire. Of the oldest group, 100 (81%) also underwent a clinical examination focusing on occlusal factors and function and dysfunction of the masticatory system. For analyses of predictors of some oral parafunctions and tooth wear registered at the 20-year follow-up, logistic regression was used with recordings at the first examination as independent variables. Results: Subjective reports in childhood of bruxism (defined as tooth clenching during daytime and/or tooth grinding at night), clenching only, grinding at night only, nail biting, and/or other parafunctions were predictors of the same oral parafunctions 20 years later. There were different predictors of the 2 components of bruxism, daytime tooth clenching and tooth grinding at night. Postnormal occlusion (Angle Class II malocclusion) and tooth wear in childhood predicted increased tooth wear in adulthood. Subjects with nonworking-side interference had less anterior tooth wear than those without such interference. Conclusion: Oral parafunctions in childhood may be a persistent trait in many subjects. Postnormal occlusion and tooth wear in childhood predicted increased anterior tooth wear 20 years later, whereas nonworking-side interference reduced the risk for such wear in 35-year-old subjects. Keywords: dental occlusion, longitudinal study, nail biting, tooth clenching, tooth grinding