In New Zealand, as elsewhere, caries prevalence has declined since the 1950s; this has been accompanied by a change in the intra-oral pattern of the disease. This is illustrated by analysis of data for 12-year-old children. However, because treatment services for children in New Zealand are so comprehensive, the DMF index is primarily a count of restorations placed. This treatment overlay can distort the true caries prevalence and has been a confounding factor in assessment of the change in caries over time. Measurement of the fine gradations of ongoing change in the present low-caries-prevalence population requires the use of a more sensitive indicator than the DMF indices. When the timing of various forms of fluoride supplementation is correlated with the decline in caries, the decline continues beyond the time of maximum population coverage with fluoridated water and fluoridated toothpaste. Thus an explanation of the convergance of caries prevalence in fluoridated and non-fluoridated areas since the 1970s may require a re-assessment of the fluoride effect. This convergence, and the overall decline during the last decade without known additional fluoride supplementation, suggest that factors other than fluoride, such as food additives and antibiotics, may have contributed.