An immunoglobulin G (IgG) antibody capture assay (GACRIA) methodologically distinct from other assays for the detection of antibodies to human T-lymphotropic virus type III/lymphadenopathy-associated virus (anti-HTLV III/LAV) was developed and used to test 1,055 serum samples previously screened by a competitive assay (COMPRIA). Eight hundred and twenty-three (78%) sera were positive and 129 (12.2%) negative in both assays. The coefficient of correlation between the two assays was 0.87, and COMPRIA appeared more sensitive than GACRIA. On retesting 103 sera that gave initially conflicting results, 81 discrepancies were resolved, 77 because of a change in the COMPRIA result. The 22 persistently discrepant samples gave an equivocal or positive result by COMPRIA and were negative by GACRIA. Thirteen of these 22 were positive in a commercial ELISA (ELAVIA). Twenty of them were also tested by Western blot; all were reactive with at least two HTLV III/LAV protein bands. The persistently discrepant samples were of four types: laboratory control sera (n = 2); naturally occurring weakly reactive samples (n = 5); samples that were anti-HTLV III/LAV IgG negative but IgM positive (n = 5; all five individuals from whom these samples were drawn were symptomatic); samples from one laboratory that were probably accidentally contaminated with anti-HTLV III/LAV-positive serum (n = 10). It was concluded that GACRIA for anti-HTLV III/LAV is specific and adequately sensitive and, in conjunction with other assays, is a useful confirmatory test whose format could readily be changed to ELISA.