Objectives This study prospectively examines the relationship of possible aetiological factors to the presence of tooth erosion in a cohort of children. Patients and methods A random sample of 1,753 children was examined at age 12 and 1,308 of the same children were re-examined at age 14 years. The children were asked to complete questionnaires on both occasions, 1,149 subjects gave usable replies. The erosion index used was based upon the 1993 Survey of Children's Dental Health. Results were analysed using logistic regression. Results At age 12 significant positive associations were found between erosion and decay experience (odds ratio [OR]=1.48), drinking fruit juice (OR=1.42) or fizzy pop (OR = 1.59-2.52, depending on amount and frequency). The presence of calculus (OR 0.48) or eating fruit other than apples or citrus fruit (OR 0.48) reduced the chances of erosion. High consumption of carbonated drinks increased the odds of erosion being present at 12 years by 252% and was a strong predictor of the amount of erosion found at age 14. Conclusions Of the factors investigated, a history of dental caries and a high consumption of carbonated drinks were most closely related to the presence of dental erosion. The risk of erosion bore a strong relationship to the amount and frequency of carbonated drink consumption.It is only relatively recently that tooth erosion has been recognized as presenting a dental health problem in children. In 1993, the prevalence of erosion was reported as part of the national survey of Children's Dental Health in England and Wales 1 for the first time. Fifty per cent of 6-year-olds were found to have erosion and in 23% of children it had progressed into dentine. Thirty two per cent of 14-year-old children had erosion of the permanent dentition, with dentinal involvement present in 2%.The three main types of non-carious loss of dental hard tissue are attrition, abrasion and erosion. 2 Attrition implies removal of tooth surface due to movement of teeth against one another, pos-