Acetylcholine receptor (AChR) binding and AChR modulating antibodies were found with approximately the same frequency (86%) in 349 patients with myasthenia gravis (MG). However, the total yield of positive serological results was significantly improved (90%) by assaying AChR modulating antibodies when AChR binding antibodies were not detected, because in 27 patients (8%) only one of the two tests was positive. The immunoprecipitation test for AChR blocking antibodies yielded fewer positive results (52%), but there was a significant correlation between the degree of AChR blockade and generalization of muscle weakness. In no patient was this the only positive test, because the test for AChR modulating antibodies in this study detected both AChR blocking and modulating antibodies. Human muscle AChR was used as antigen in all tests. False positive results were rare and were attributed to unexplained antibodies to 125I-alpha-Bgt (AChR binding antibody assay) and recent general anesthesia using muscle relaxants (AChR blocking and AChR modulating antibody assays). Unexplained positive results, documented previously in 5% of patients with the Lambert-Eaton myasthenic syndrome and amyotrophic lateral sclerosis (ALS), were found in this study in two of 22 patients with ALS, but in none of 427 patients with miscellaneous neurological diseases. Patients with severe generalized MG and/or thymoma tended to have higher titers of AChR binding antibodies and greater AChR modulating antibody activity. However, some patients with severe muscle weakness had low titers of antibodies, and some patients in remission or with only ocular manifestations had high titers. These seemingly paradoxical results reflect heterogeneity in the specificities, affinities, and isotypes of anti-AChR antibodies. To effect pathogenicity, antibodies must have access in vivo to extracellular antigenic sites on the AChR. One would anticipate that antibodies with greatest pathophysiological potential would be of an IgG with greatest pathophysiological potential would be of an IgG subclass that activates complement, or of a specificity that competes for acetylcholine's binding site on the receptor or readily cross-links two AChR molecules to trigger receptor modulation (eg, by binding to sites on the duplicated alpha-subunit). In patients with suspected MG who lack serological evidence of anti-AChR antibodies, motor endplate biopsy is required for microelectrophysiological, immunochemical, and ultrastructural studies to establish with certainty whether or not the condition is acquired MG.