SUMMARY.We have compared two different second‐generation (2.O) enzyme‐linked immunosorbent assays (ELISA) for the presence of antibodies to hepatitis C virus (anti‐HCV) in blood from volunteer, unpaid donors. At two separate blood centres, a total of 21431 donor samples were tested with Abbott Anti‐HCV 2.0 ELISA and Ortho Anti‐HCV 2.0 ELISA. Samples found to be repeatedly reactive were tested by supple‐mental/investigational assays, MATRIX HCV (Abbott) and anti‐HCV RIBA II (Ortho/Chiron), to ‘confirm’ the presence of anti‐HCV. Discordant ELISA samples were additionally tested by the polymerase chain reaction (PCR) for the presence of HCV RNA. The Abbott anti‐HCV assay had a repeatedly reactive rate of 0.59% (127/21431) and the Ortho anti‐HCV assay 0.51% (110/21431). Overall agreement between assays was 99.76%. 72/127 (56.7%) of Abbott repeatedly reactive samples confirmed on MATRIX and 61/127 (48.0%) on RIBAII; 70/110 (63.6%) of Ortho repeatedly reactivate samples confirmed on MATRIX and 61/110 (55.5%) on RIBA II. Discordant ELISA samples tested by PCR yielded negative results. Hence the two ELISA had equal sensitivity, as defined by detection of true positive samples: the slightly lower specificity of the Abbott Anti‐HCV 2.0 ELISA may be owing to culling of donors with a false positive test by Ortho's Anti‐HCV 1.0 and 2.0 ELISA tests (the routine tests in place at each blood centre). A sample found to be repeatedly reactive by two different ELISA tests for anti‐HCV is likely to be a true positive and may not require f