Free plasma thyroxine index (FT4I) and plasma concentrations of triiodothyronine (T3) were routinely determined on admission in 1862 hospital patients without signs or history of thyroid dysfunction. Total thyroxine (T4) and FT4I values were beyond confidence limits in 16.3% and 7.5% respectively of the patients, whereas T3 values were low in 23%. Among the 84 patients with high FT4I, 31 had clinically unsuspected hyperthyroidism confirmed by the TRH test (T3 increased in 66% of the cases). Among the 46 patients with low FT4I, 14 had demonstrable hypothyroidism (low T3 in 65% of the cases). In this population, the prevalence of proven thyroid dysfunction without clinical symptoms was 1.66% for hyperthyroidism and 0.75% for hypothyroidism. The diagnostic value of normal FT4I was estimated at 50%. The persistence of an abnormal FT4I on a second determination indicated the presence of hyperthyroidism in 72% of patients with high FT4I values and of hypothyroidism in 74% of patients with low FT4I values. In doubtful cases, TSH assays or TRH tests led to the concept of "transient" dysthyroidism, and the potential total prevalence of routinely discovered dysthyroidism could be estimated at 3.54% including 2.47% for hyperthyroidism. The latter occurred in 95% of people older than 50, with a sex ratio of 0.94. The cost of diagnosis for each new case clinically unsuspected hyperthyroidism is 1200 Z, but this could be reduced to 450 Z if only patients over 50 years of age were investigated and if FT4I determinations were replaced by free T4 determinations.