The serum agglutination test (SAT) and 2-mercaptoethanol (2ME) agglutination were used in studies of the sera of 214 patients in whom brucellosis was suspected. On the basis of historical, epidemiologic, clinical, and serologic data, four groups were identified: group I (108 cases) had negative agglutination reactions, and brucellosis was considered unlikely; group II (57 cases) had positive agglutination reactions, and active brucellosis was diagnosed; group III (37 cases) had positive agglutination reactions, but other factors--notably, a history of prior infection--made inactive brucellosis likely; and group IV (12 cases) had positive agglutination reactions, but insufficient data were available for further classification. Most patients with active brucellosis had agglutinin titers of greater than or equal to 160; however, no single titer was always diagnostic. Although more sensitive tests are available, agglutination reactions provide data sufficient to differentiate active from inactive disease when other factors are considered and follow-up sera are tested. This article discusses individual cases and reviews the literature on the diagnosis of brucellosis.