A commercial-latex agglutination (LA) assay (Rubascan®, Hynson, Westcott and Dunning, Baltimore, Md) performed on undiluted serum may be more sensitive than is the standard hemagglutination-inhibition assay for the detection of rubella antibody. To investigate whether a positive result of an LA test on undiluted serum was correlated with rubella immunity, we identified 29 persons whose sera produced negative results when initially tested by LA at a 1:10 dilution; 15 of these sera produced positive results and 14 produced negative results when retested undiluted. The immune status of each person for rubella was determined with a test for IgM rubella antibody on serum samples obtained three to five weeks after rubella immunization. None of the 15 persons whose preimmunization sera produced positive results by LA when tested undiluted developed IgM rubella antibody, as compared with 10 of the 14 persons whose preimmunization sera produced negative results (P <0.001). Data from this study support the specificity of a positive result of an LA test performed on undiluted serum, but a 1:5 dilution may be optimal for determination of rubella immune status.