A prospective population study of women in Gothenburg, Sweden was started in 1968-69 and comprised 1462 women aged 38, 46, 50, 54, or 60 years at baseline. Follow-up studies were carried out in 1974-75, 1980-81, and 1992-93. The baseline study included an extensive medical and dental examination. Serum mercury concentration (beta-HG) was determined in deep-frozen samples from all participants in 1968-69 and in a random subsample of sera from participants in 1980-81, about 20 years after the baseline examination. S-Hg was statistically significantly correlated with number of amalgam fillings at both examinations. Of 30 defined symptoms and 4 different clusters of symptoms, no one was independently correlated with S-Hg measured in the samples from 1968-69, while there was a negative statistically significant correlation with over-exertion and poor appetite in 1980-81. Blood hemoglobin and serum B-12 concentrations in 1968-69 were statistically significantly and positively correlated with S-Hg, while erythrocyte sedimentation rate and the serum concentrations of potassium and triglycerides were significantly and negatively correlated with S-Hg, also after including potential confounders. Blood hematocrit examined in 1980-81 was negatively correlated with S-Hg. When including potential confounders, serum IgA was also statistically significantly correlated with S-Hg, but not in univariate analysis. No statistically significant correlation was observed between S-Hg, on the one hand, and the incidence of diabetes, myocardial infarction, stroke, or cancer on the other, while a statistically significant negative correlation was observed with overall mortality when age and education were included as background variables. There were some correlations between biological variables and S-Hg, probably of no negative clinical significance, and we conclude that there is no association between disease and S-Hg on a population basis in middle-aged and older women.