The serum T₃ to T₄ ratio is a useful indicator for differentiating destruction-induced thyrotoxicosis from Graves' thyrotoxicosis. However, the usefulness of the serum free T₃ (FT ₃) to free T₄ (FT₄) ratio is controversial. We therefore systematically evaluated the usefulness of this ratio, based on measurements made using two widely available commercial kits in two hospitals. Eighty-two untreated patients with thyrotoxicosis (48 patients with Graves' disease and 34 patients with painless thyroiditis) were examined in Kuma Hospital, and 218 patients (126 with Graves' disease and 92 with painless thyroiditis) and 66 normal controls were examined in Ito Hospital. The FT₃ and FT₄ values, as well as the FT₃/FT₄ ratios, were significantly higher in the patients with Graves' disease than in those with painless thyroiditis in both hospitals, but considerable overlap between the two disorders was observed. Receiver operating characteristic (ROC) curves for the FT₃ and FT₄ values and the FT₃/FT₄ ratios of patients with Graves' disease and those with painless thyroiditis seen in both hospitals were prepared, and the area under the curves (AUC), the cut-off points for discriminating Graves' disease from painless thyroiditis, the sensitivity, and the specificity were calculated. AUC and sensitivity of the FT₃/FT₄ ratio were smaller than those of FT₃ and FT₄ in both hospitals. The patients treated at Ito hospital were then divided into 4 groups according to their FT₄ levels (A: ≤2.3, B: >2.3~≤3.9, C: 3.9~≤5.4, D: >5.4 ng/dl), and the AUC, cut-off points, sensitivity, and specificity of the FT₃/FT₄ ratios were calculated. The AUC and sensitivity of each group increased with the FT₄ levels (AUC: 57.8%, 72.1%, 91.1%, and 93.4%, respectively; sensitivity: 62.6%, 50.0%, 77.8%, and 97.0%, respectively). The means ± SE of the FT₃/FT₄ ratio in the Graves' disease groups were 3.1 ± 0.22, 3.1 ± 0.09, 3.2 ± 0.06, and 3.1 ± 0.07, respectively, versus 2.9 ± 0.1, 2.6 ± 0.07, 2.5 ± 0.12, and 2.3 ± 0.15, respectively, in the painless thyroiditis groups. In the painless thyroiditis patients, the difference in the FT₃/FT₄ ratio between group A and group D was significant (p<0.05). Thus, the FT₃/FT₄ ratio in patients with Graves' disease likely remains unchanged as the FT₄ level rises, whereas this ratio decreases as the FT₄ level rises in patients with painless thyroiditis. In conclusion, the FT₃/FT ₄ ratios of patients with painless thyroiditis overlapped with those of patients with Graves' disease. However, this ratio was useful for differentiating between these two disorders when the FT₄ values were high.