"Sandwich"-type assays are subject to positive interference by the patient's "heterophile" antibodies. If present, these bind to the animal immunoglobulins in the assay reagents, forming artefactual sandwiches indistinguishable from those formed with the analyte itself. Immunoglobulins from non-immunized animals, added to the assay reagents, can diminish this effect by blocking the patient's antibodies. Elsewhere, we studied several patients with anti-mouse immunoglobulin activity, whose serum gave spuriously high results for thyrotropin (TSH) concentrations. Here we have studied this phenomenon by adding, to pooled zero-TSH serum, antibodies to mouse, goat, and horse immunoglobulins and then assaying TSH by several other sandwich-type assays involving mouse monoclonal antibodies. Assays not supplemented with blocking immunoglobulins from mice or other animals were more susceptible to this effect. When large amounts of antibody were added, the antibody excess diminished the interference. However, the presence of blocking immunoglobulins could reverse such antibody excess, actually enhancing, instead of diminishing, the positive interference. Users should be aware that blocking immunoglobulins may diminish but not necessarily eliminate this problem with such assays.