Although toxicological and metabolic interactions of arsenic (As) and selenium (Se) have been suggested by epidemiolgical literatures, the past experimental studies mostly focused on acute, high-dose interaction, leaving the long-term, low-dose interaction unexplored. In the present study pregnant mice, fed either Se-deficient or adequate (0 or 5 micromol Se/kg diet, respectively) diet, were given oral gavage of sodium arsenite (0 or 58 micromol/kg per day; chosen as less than half of the fetotoxic dose in this protocol) from gestational day (GD) 7-16. The levels of As and Se as well as five selenoenzymes (glutathione peroxidase (GPx), thioredoxin reductase (TRxR), and type-I, -II and -III iodothyronine deiodinases (DI-I, -II and -III) were examined on GD17 in the tissues of dams and of fetus. The Se-deficient mice showed significantly enhanced accumulation of As compared to the Se-adequate mice in maternal liver (increased by 48%) and fetal brain (by 31%). Although no direct evidence of the enhanced toxicity in the Se-deficient group was obtained, the As exposure affected the levels of Se and selenoenzymes, an effect which was more discernible in Se-deficient group. Although most of theses changes were mild or moderate, the DI-II activity in Se-deficient fetal brain showed a drastic four-fold increase by As exposure, suggesting a possible disturbance of thyroid hormone environment in the fetus. These data suggested that apparently non-toxic, in utero dose of As, showed enhanced accumulation when combined with Se-deficiency and could affect the metabolism/kinetics of Se in fetal brain, which might result in developmental toxicity in mice.