We describe a clinically euthyroid 60-year old woman with a past history of 1311 therapy for treatment of possible toxic nodular goitre. She had an elevated thyroxine level of 188 microgram/L (normal range 50-125 microgram/L) following her therapy, and her TSH was within normal limits at 4.7 mU/L. However her T3 level, as determined by an RIA technique employing charcoal to separate bound and free T3, was not measurable. T3 added to the patient's serum could not be recovered, therefore the presence of an unusual protein capable of binding T3 was suspected. To avoid this interference, T3 analysis was performed on an ethanol extract of the patient's serum and was found to be 17 microgram/L (normal range 0.8-1.8 microgram/L). At this time her thyroid microsomal antibody titre was 1:100,000. A protein, capable of binding more than 70% of the patient's T3, was demonstrated in the gamma globulin fraction by agarose gel electrophoresis. This protein did not bind T4. Scatchard analysis for T3 binding revealed a protein, presumably IGG, with a binding affinity of 2 x 109 L/mole and binding capacity of 50 microgram/L. This case exemplifies the caution that must be taken in interpreting thyroid function tests. When thyroid hormone levels are inappropriate to the clinical status of the patient the presence of circulating antibodies which can bind the thyroid hormones should be considered.