A 4-yr mixed-longitudinal study to determine the prevalence of caries in 7-13-yr-old Tanzanian children was started in 1984. The parameters considered were age, locality, Socio-Economic Status, and sex. Locality was composed of urban (Morogoro town), rural (Morogoro District), and rural areas in the District with an average fluoride level of 0.5 ppm or more in all drinking water present. SES was established based on the occupation of the father or mother and on housing conditions. Overall, the reproducibility of the dental conditions studied (D2MT/S and D3MT/S) was high, with lower scores for the conditions including early enamel lesions (D3MT/S). The reproducibility of the SES scoring system was high (kappa = 0.96 and kappa = 0.90), but the association over the 2 yr of measurement (1984 and 1988) was weak (r = 0.50). There were no restorations found. The percentage of children with caries increased with increasing age from 12-17% at age 7 to 37% at age 13. The statistical tests (ANOVA) revealed an age effect for all conditions studied in 1984, 1986, and 1988 and a locality effect in 1988 only. The mean D3MT-scores varied between 0.15 and 0.24 at age 7 to 0.76 at age 13, while the mean D3MS-scores varied between 0.27 and 0.31 at age 7 to 1.18 at age 13. In general, the caries prevalence observed was low. Children living in naturally fluoridated rural areas had significantly lower caries scores than children in non-fluoridated areas.