Dexamethasone, 2 mg every 6 hours for 4 doses, was given to 4 hypothyroid patients receiving treatment with synthetic thyroxine (T4) and to 8 untreated hyperthyroid patients with Graves' disease, and serum concentrations of thyroid hormones were measured by radioimmunoassays. Serum concentration of 3,3'5'-triiodothyronine (reverse T3, rT3) increased appreciably within 8 hours after the first dose of dexamethasone, was maximum at 24-32 hours after beginning dexamethasone, and remained elevated for about 24 hours after discontinuing the steroid. The mean baseline serum rT3 was 58 ng/per 100 ml in treated hypothyroid patients and 119 ng per 100 ml in patients with Graves' disease; the corresponding maximal post-dexamethasone serum rT3 values were 87 and 170 serum concentration of 3,3',5-triiodothyronine (T3) decreased. The decrease in serum T3 was significant at about 24 hours after beginning dexamethasone and was maximal at about 30 hours in both groups of cases under study. The decrease in serum T3 persisted in treated hypothyroid cases for about 24-48 hours and in Graves' disease cases as long as studied, at least 5 days after discontinuing hexamethasone. The changes in serum rT3 and T3 could not be attributed to the effect of dexamethasone on serum protein binding of the iodothyronines because the dialyzable fractions of rT3 and T3 following steroid administration were not different from those before it. Serum T4 did not change appreciably in treated hypothyroid cases, but decreased in Graves' disease cases from a mean baseline value of 23.5 mug per 100 ml to 18.4 mug per 100 ml 3 days after beginning dexamethasone. In addition, 3 hyperthyroid cases were studied before, during, and after administration of dexamethasone, 2 mg every 6 h for 5 days. Serum rT3 increased again as noted above and the increase persisted until about 24 hours after the last dose of the steroid. Serum T3 decreased considerably and remained decreased as long as studied, at least 4 days after discontinuing the steroid. Serum T4 decreased appreciably in 2 of the 3 cases studied. The data suggest that 1) conversion of T4 to T3 and to rT3 may occur via two distinct pathways in the metabolism of T4; 2) the changes in serum rT3 and T3 observed in our study may be due in part at least to a steroid-induced 'shift' in the metabolism of T4 whereby conversion of T4 to T3 is diminished and that to rT3 is enhanced; 3) in addition to the effect on peripheral metabolism of T4, steroids appear to reduce the circulating thyroid hormones in Graves' disease by another mechanism, probably by reduction in thyroid secretion.