The question of the teratogenicity of lithium carbonate was tested in two inbred strains of mice that are susceptible to teratogens. Strains 129 and A/J mice were treated with varying dosages of lithium carbonate on day 8, 9, or 10, and strain A/J mice were sequentially treated on days 11, 12, and 13 of gestation. Another group of mice of strain 129 were given lithium carbonate in drinking water throughout pregnancy. Levels of lithium were determined in mouse serum (and in brain) of mothers and their offspring with the purpose of comparing equivalent values in mice and in man. A baseline dosage of 0.8 mg of lithium carbonate was chosen, since this dose produced a serum lithium level of 0.8 meq/L, which is comparable to values of patients treated for manic-depressive illness. This baseline dosage or 2 and 4 times this amount did not prove teratogenic in strain 129 mice. However, 5.0 mg (200 mg/kg), which is one-half the LD50 value, caused 41% malformations. The baseline dosage did not cause malformations in strain A/J when given on day 11, 12, or 13 of gestation. Lithium carbonate given in the drinking water to strain 129 mice throughout pregnancy at a concentration that produces serum levels in mice consistent with human values was deleterious. The data suggest that whole or part of total litters were eliminated. The few survivors were apparently normal. Our results suggest that 6 times the therapeutic serum lithium level in humans is teratogenic in mice, as tested by acute experiments. Chronic treatment with a quantity of lithium that produces a serum level equivalent to that of human values is toxic to whole or part of the litters of mice.