Three groups of female monkeys (Macaca fascicularis) were exposed to methylmercury (MeHg, p.o. 50 micrograms Hg/kg body wt per day) for 6, 12, or 18 months. One group was exposed to MeHg for 12 months and kept unexposed for 6 months before sacrifice. Another group of three monkeys was exposed to HgCl2 i.v. for 3 months. Total and inorganic mercury concentrations in occipital pole and thalamus were determined by cold vapor atomic absorption spectroscopy. Selenium concentrations were analyzed by hydride generation atomic absorption spectroscopy. The results indicated an association between concentrations of inorganic mercury and selenium in both occipital pole and thalamus in the MeHg-exposed animals. A linear regression model using concentrations of inorganic mercury (nmol/g wet wt) as independent variable, and selenium concentrations (nmol/g wet wt) as the dependent variable showed significant correlations between the variables in both occipital pole and thalamus (r = 0.85 and r = 0.91, P < 0.0001). The intercept of the regression line was slightly lower (about 2 nmol Se/g wet wt) than the selenium concentrations found in control monkeys (about 3 nmol Se/g wet wt). There was a tendency to a "hockey stick"-shaped relationship between concentrations of selenium and inorganic mercury in the thalamus of monkeys with ongoing exposure to MeHg. An important role for selenium in the retention of mercury in brain is indicated.