Mild dental fluorosis is frequently linked to fluoridated water, but discretionary fluoride sources may also be important. The aim of this study was to record age of weaning and fluoride exposure from water, toothpaste and supplements, and to relate these to the presence of caries and fluorosis in children born in 1983. In Perth (Western Australia) 14 school classes were selected. The 350 children (mean age 7.5 years) ultimately included gave fluoride exposure data for the period birth to 4 years of age. Caries (DMFT, WHO criteria, no radiographs) and dental fluorosis (TF index, dry permanent incisors) were registered clinically. Most (89%) children had lived at least 2.5 years in a fluoridated area. Supplement use was minimal and unrelated to caries or fluorosis. Mean age of weaning of those who had been breast-fed was 7.7 months; by 9 months, 74% had been weaned. Eighty-five percent liked toothpaste, 60.7% had swallowed it, and the mean age of starting to use it was 1.5 (SD 0.96) years. Caries prevalence was 0.1 and mean DMFT was 0.13. The prevalence of fluorosis was 0.48; 63% of fluorosis was TF score 1. Residence in a fluoridated area for > or = 2.5 of the first 4 years of life had an odds ratio (OR) of 4.9 for fluorosis. Weaning before 9 months of age, swallowing toothpaste and liking toothpaste were also statistically significant risk factors. Major risk factors for more severe fluorosis (TF > or = 2) were early weaning and swallowing toothpaste (ORs 2.77 and 2.64, respectively). Residence in a fluoridated area (OR 2.2) was not a statistically significant risk factor.(ABSTRACT TRUNCATED AT 250 WORDS)