CLINICAL diagnosis of human brucellosis is often difficult, especially in chronic and subclinical infections, in which the symptoms may be relatively non-specific. The diagnosis must often be made on the basis of laboratory tests; and, because brucellosis is a possible diagnosis in patients with pyrexia of undetermined origin, the help of the laboratory is often sought. Brucella abortus is isolated from only a minority of infected patients (Wilson and Merrifield, 1951) and most of these have an acute infection. Most commonly, therefore, the diagnosis is made by serological methods particularly in the chronic form of the disease. Several tests are usually employed in attempts to detect and distinguish acute, chronic and past infections. These methods include the direct and mercaptoethanol agglutination tests, a complement-fixation test and an anti-human-globulin (Coombs) test (Kerr et al., 1968). The results of these tests are not always easy to interpret because none is specific for a single immunoglobulin class, and several tests are required for what is often merely a screening procedure. Parratt et al. (1977) introduced a radioimmunoassay (RIA) to measure Br. abortus-specific antibodies of the IgG, IgM or IgA class. This specificity for antibody class simplifies interpretation and the test can demonstrate small amounts of antibody. RIA would be an adequate substitute for the current range of tests but its use is restricted by the availability of facilities for handling and counting 135I. A mixed reverse passive antiglobulin haemagglutination technique for brucella serology has been reported by Coombs et al. (1978); this, too, is antibody-class specific and may be more suited to a routine diagnostic laboratory than RIA. Enzyme-labelled immunosorbent assay (ELISA) is comparable in sensitivity to RIA and has been found useful in many areas of serology (Voller, Bartlett and Bidwell, 1978). ELISA techniques for detecting Br. abortus antibody in cattle (Saunders et al., 1977) and in rabbit sera (Carlsson, Hurvell and Lindberg, 1976) have been described. In both methods, culture of the organism and extraction of the antigen were necessary. The present investigation was undertaken to study the application of ELISA in brucella serology with a more convenient antigen.