Summary The StaT4 test is a technically simple test of thyroxine measurement, which apparently gives results in good agreement with those derived by radioimmunoassay in most patients. In pregnancy, however, falsely low serum thyroxine values were found with the StaT4 test when compared with either a radioimmunoassay or the Thyopac-4 test. Detailed studies showed that the discrepancy resulted from the extraction technique used in the StaT4 test, which appears to allow the patient's thyroxine binding proteins to remain in the assay system. Similar method discrepancies were observed in other clinical conditions associated with elevated thyroxine binding globulin concentration. This study, as well as being of practical clinical importance, has implications for the design of competitive protein binding tests of thyroid function.