National Institute of Preventive Medicine (NIPM) has succeeded in the development of an enzyme immunoassay (EIA) kit for detection of hepatitis B surface antigen. The sandwich principle was used for the test. Guinea pig anti-HBs IgG was used for coating microtiter plates and Horseradish peroxidase was conjugated with goat specific anti-HBs. Its stability is longer than 4 months. The lowest detectable dose is 0.7 ng/ml or better for subtype ad of HBsAg tested with Hepatitis Sensitivity Panel, Bureau of Drug and Food, Department of Health. The regression curve was determined by testing 66 samples with Auszyme II (EIA Kit. Abbott Lab.) and NIPM Kit, while Auszyme II used as a reference kit. The two EIA kits correlated well with a coefficient of determination (r2) of 0.86. Evaluation on 1,157 patients' and officers' serum samples in Tri-Service General Hospital showed that the positive rate was 24.7% (286/1,157) by RIA (Clinical Assays, Travenol Lab., USA) and that of NIPM EIA Kit was 24.4% (282/1,157). There was no statistical significance in terms of positive rate (p greater than 0.05). The positive rates of 534 blood donors are 22.1% (118/534) and 21.9% (117/534) respectively. Another evaluation on 974 serum samples in National Taiwan University Hospital showed that the positive rate was 27.2% (265/974) by Ausria II-125 (RIA. Abbott Lab.) and that of NIPM EIA Kit was 27.4% (267/974). The undetectable rate and false positive rate of NIPM EIA Kit were 0.41% (4/974) and 0.62% (6/974) respectively. In comparison with four other kind of commercial EIA Kits, the results of NIPM EIA Kit were satisfactory also. We have scaled up the reagent preparations for the kit, except the antibody coated microtiter plate preparation. At the end of March 1985 we will supply 850 EIA kits for the Development Center for Biotechnology for their hepatitis B vaccine production program.