Commercially available kits for detection of hepatitis B surface antigen (HBsAg) and hepatitis B surface antibody (anti-HBs) by enzyme immunoassay (EIA) were evaluated in American Samoa during a public health programme to eliminate the transmission of hepatitis B. The first 19184 serum specimens obtained, representing 68% of the total cooperating population, were initially tested for anti-HBs, and those without detectable antibody were tested for HBsAg. All the antigen-positive serum samples, and a selection of the antigen-and antibody-negative specimens were tested by radioimmunoassay (RIA) for detection of both markers. Compared with the standard tests, the EIA kits for anti-HBs and HBsAg performed well; sensitivity and specificity were 90·3 and 96·0%, respectively, for antibody, and 97·8 and 97·9% respectively for antigen. Substantial disagreement between the EIA and RIA tests for HBsAg was found only for specimens considered weakly reactive by EIA. Few differences were found between three EIA method options for follow-up HBsAg testing of weakly reactive serum specimens; each option contributed about equally to improved test specificity for these ‘borderline’ specimens. Based on their demonstrated equivalence to the standard RIA tests, we conclude that the EIA kits for anti-HBs and HBsAg detection are suitable for use in hepatitis B control programmes in open populations.