Sera from 111 patients hospitalized on acute-care wards (including 32 in the intensive care unit) were examined for the possible presence of inhibitors of thyroxine (T4)-serum protein binding in an assay employing equilibrium dialysis. In 38 of these sera, the unbound (free) T4 fraction was 50% or more higher than the free T4 fraction in a pool of normal sera. From the free T4 fraction in each of the 111 serum samples and the free T4 fraction in the pool of normal sera, the predicted free T4 fractions in mixtures (1:1) of each of these sera with the normal pool were calculated (assuming the absence of binding inhibitors) from the appropriate mass action equations. It was reasoned that a free T4 fraction in any mixture that exceeded this predicted value would indicate the possible presence of a binding inhibitor. (The normal pool was selected for having a low serum triglyceride concentration, to minimize in vitro generation of free fatty acids.) However, for the 111 serum samples studied, the free T4 fraction in the mixture exceeded the upper 95% confidence limit of this predicted value in only one case, and then just barely. Thus, evidence for an inhibitor of T4-serum protein binding in sera from patients with nonthyroid illness could not be found. Twenty-eight of the serum samples were also examined in a similar assay that employed ultrafiltration of undiluted serum instead of equilibrium dialysis. Evidence for an inhibitor of T4-serum protein binding similarly could not be found. Because part of the reason for postulating the existence of such a binding inhibitor has been the performance of the triiodothyronine (T3) resin uptake test in patients with nonthyroid illness, an alternative explanation for this phenomenon was sought. When thyroid hormone-binding globulin (TBG) was desialylated by treatment with neuraminidase, its avidity for T4 was markedly decreased, but its avidity for T3 was unchanged. Thus, if desialylated TBG circulates in patients with nonthyroid illness as previously reported, it could explain not only the low serum T4 concentrations despite near normal immunoreactive TBG concentrations, but also the poor performance of the T3 resin uptake test (where T4 binding capacity is overestimated) in these patients.