A patient with cirrhosis of the liver who developed megaloblastosis despite normal serum and red cell folate levels after receiving conventional doses of triamterene for 2 weeks is described. Studies using short-term human bone marrow cultures indicate that triamterene interferes with the de novo deoxyribonucleic-acid-thymine synthesis from deoxyuridine and that this interference is completely corrected by reduced (folinic acid) but only partially by oxidized (pteroylglutamic acid) folate. These findings support the concept that triamterene inhibits dihydrofolate reductase activity in human bone marrow, as it does in other mammalian systems, and suggest this drug should be used with caution in patients who may have borderline folate stores (such as pregnant women and alcoholics).