: The present study aimed at investigating the nature and causes of non-parallelism in testosterone responses to serial dilutions of peripheral serum and standard LH preparations in the mouse Leydig cell in vitro bioassay of LH. Immunoadsorption with monoclonal antibody to the β-subunit of LH was used to obtain LH-free serum; the procedure removed more than 98% of the immunoassayable LH. When a constant amount of the LH-free serum was added to standard dilutions, the bioassay dose-response curves to serum dilutions and standards became parallel, i.e. the well-known source of error of this assay system was eliminated. When standard curves prepared in medium and LH-free serum (final concentration 10%) were compared, no effect of the serum was found on basal cAMP and testosterone production. However, the LH-stimulated testosterone and cAMP production were suppressed by serum by a rather constant factor of 40%. Mild heating (60°C, 15 min) or treatment with dextran-coated charcoal, but not ether extraction, was able to eliminate the inhibitory activity of the LH-free serum. Binding studies demonstrated that [125I]hCG interaction with mouse Leydig cell homogenates was inhibited by LH-free serum in a fashion indicative of reduced LH receptor number, but not of reduced binding affinity. In conclusion, these data show that human serum contains LH inhibitor(s) which affect the LH-receptor interaction and LH stimulated testosterone production in mouse Leydig cell in vitro. The effect is marked in serum concentration over 1.5% and it shows only minor variation between individual sera. This source of error can be effectively removed from the LH in vitro bioassay by using LH-free serum for preparation of dilutions of LH standards.