The water from some drilled wells in southwest Finland contains high arsenic concentrations (min-max: 17-980 microg/L). We analyzed inorganic arsenic (As-i) and organic arsenic (monomethylarsonate [MMA] and dimethylarsinate [DMA]) species in urine and conducted a clinical examination of current users (n = 35) and ex-users (n = 12) of such wells. Ex-users had ceased to use the water from the wells 2-4 months previously. Urinary arsenic species were also analyzed from persons whose drinking water contained less than 1 microg/L of arsenic (controls, n = 9). The geometric means of the concentrations of total arsenic in urine were 58 microg/L for current users, 17 microg/L for ex-users, and 5 microg/L for controls. The excreted arsenic was associated with the calculated arsenic doses, and on average 63% of the ingested arsenic dose was excreted in urine. The ratios of MMA/DMA and As-i/As-tot (As-tot = As-i + MMA + DMA) in urine tended to be lower among the current users and in the higher exposure levels than in controls, suggesting that As-i was better methylated in current users. However, the differences were mainly explained by age; older persons were better methylators of inorganic arsenic than younger individuals. The arsenic content of hair correlated well with the past and chronic arsenic exposure; an increase of 10 microg/L in the arsenic concentration of the drinking water or an increase of 10-20 microg/day of the arsenic exposure corresponded to a 0.1 mg/kg increase in hair arsenic. The individuals were interviewed and complained of muscle cramps, mainly in the legs, and this was associated with elevated arsenic exposure. The present study demonstrates that arsenic methylation has no threshold at these exposure levels.