Higher positivity rates for prevalence of anti-HTLV-I antibody have been reported for the gelatin particle agglutination (PA) assay when compared to that of the indirect immunofluorescence assay using acetone-fixed HTLV-I producing cells (IF-FA). To evaluate the discrepancy between these two screening methods, PA-positive/IF-FA-negative sera were tested by four additional assays for anti-HTLV-I: indirect immunofluorescence assay using live HUT102 cell membranes (IF-MA), enzyme immunoassay (EIA), radioimmunoprecipitation (RIP), and Western blotting (WB). Sera obtained from 6915 Japanese blood donors were assayed for anti-HTLV-I antibody by PA, and 389 (5.6%) were positive. These 389 sera were re-examined by IF-FA, and 29 (7.5%) were negative. Sufficient material was present for 20 of the 29 PA-positive/IF-FA-negative sera for further evaluation by the IF-MA, EIA, RIP, and WB. Fifteen (75%) of the 20 were positive by IF-MA, but only 6 (30%) were positive by EIA. Both RIP and WB confirmed 17 (85%) of the samples, with each detecting a serum that was negative by the other. Thus, 18 (90%) of the 20 were confirmed by either RIP or WB. The nonconfirmed sera were all positive on PA at low titer. These findings suggest that the PA assay is more sensitive than either IF-FA or EIA.