Manganese exposure and biomarker concentrations during early pregnancy and lactation were investigated in 408 women living in an area with elevated concentrations of both arsenic and manganese in drinking water derived from wells. About 40% of the water samples had manganese concentrations above the World Health Organization's guideline value and showed a strong inverse correlation with arsenic concentrations. Water manganese was found to correlate to urine concentrations, but not to blood or breast milk concentrations. No correlations were found among manganese concentrations in urine, blood, or breast milk. Compared to other populations, manganese concentrations in both urine and blood, but not breast milk, were elevated in the Bangladeshi women and more similar to those of occupationally exposed groups. The lack of associations with water manganese is likely due to variable exposure via water and food, and differences in bioavailability, as well as a complex and/or strict regulation of intestinal manganese absorption, in turn being influenced by nutritional as well as physiological and genetic factors. The results indicate that elevated maternal manganese exposure does not necessarily lead to exposure of breast-fed infants, stressing the importance of breast feeding in high manganese areas. However, the implications of fetal exposurefrom elevated maternal exposure need further investigation.