Hormonal production of the thyroid gland is constituted of thyroxine or T4 (80%) and triiodothyronine or T3 (20%). In the circulation, whole T4 originates from thyroid secretion but most of T3 (80%) is produced extrathyroidally from T4 deiodination. Conversion of T4 to T3 may be influenced by various conditions and circulating T3 is a less reliable reflection of thyroid hormone production than T4. In serum most of T4 and T3 is bound to binding proteins and only 0.02% of T4 and 0.3% of T3 is free. Because of their higher diagnostic performance, free T4 (FT4) and free T3 (FT3) measurements have superseded total (free + bound) hormone determination. Total hormone measurements remain useful for research studies or in case of severe hyperthyroidism. Equilibrium dialysis/RIA is considered as the reference method for free hormone measurements. Routine clinical laboratories use automated direct two-step or one-step immunoassays with a high molecular weight ligand or labelled antibody. Free hormone measurement remains technically demanding, especially in sera from severe non-thyroid ill patients with low serum thyroxine binding capacity. Interference from anti-thyroid hormone antibodies and familial dysalbuminemic hyperthyroxinemia depends on the assay method, but is now less marked and less frequently detected. To be able to correctly interpret the results of an assay, it is necessary to assess its performance in biologically and clinically well-characterised serum samples. FT4, and FT3 measurements, if FT4 is normal and hyperthyroidism suspected, are used to confirm and assess the level of hypo and hyperthyroidism (overt or subclinical). When the thyroidal status is unstable (first months of a thyroid treatment, altered L-T4 dose, subacute thyroiditis) or when the hypothalamic-pituitary function is disturbed (central hypothyroidism), TSH determination is diagnostically misleading and only free hormone measurements are reliable for thyroid function assessment.