Testicular endocrine function was assessed in 54 cryptorchid boys, 9 of whom had previously been treated with human chorionic gonadotrophin (HCG). Two different HCG tests were used, the first being short (5,000 IU/m2) and the second more prolonged (3 x 1,500 IU). The testosterone response to the two tests is both quantitatively and qualitatively different. HCG therapy abolishes the differences between cryptorchid boys and controls, which may explain the conflicting results of HCG tests in literature. The results also suggest that the absence of the late testosterone response during the short test in the cryptorchid boys not treated with HCG is due to a reduced and exhaustable Leydig cell function.