By using trypsin-treated human type O cells as indicators, we compared the abilities of four polyanion-divalent cation combinations (heparin-MnCl(2); high-and low-molecular-weight dextran sulfate-CaCl(2); and sodium polyanetholesulfonate [SPS]-CaCl(2)) for removal of serum non-immunoglobulin (lipoprotein) inhibitors of rubella hemagglutination. The combination of SPS-CaCl(2) was found to be the most effective, precipitating completely the pre-beta and beta-lipoproteins and reducing the alpha-lipoprotein levels by more than 50%. Hemagglutination patterns after this treatment were clear and stable, and, when normal sera were tested, hemagglutination-inhibition (HI) titers were comparable to those obtained after standard heparin-MnCl(2) treatment. High-molecular-weight dextran sulfate-CaCl(2) removed serum lipoproteins almost as effectively as SPS-CaCl(2). However, problems of nonspecific agglutination and the heavy hemagglutination patterns resulting made this combination unacceptable for routine purposes. Neither low-molecular-weight dextran sulfate-CaCl(2) nor heparin-MnCl(2) removed the pre-beta lipoproteins completely, and occasionally traces of beta-lipoprotein also remained after treatment. The presence of pre-beta lipoproteins in normal sera after treatment may be of no consequence in the HI test since we have found that the very-low-density lipoprotein fractions obtained by ultracentrifugal methods from normal sera (those corresponding to the pre-beta fractions obtained by electrophoresis) had no HI activity. However, very-low-density lipoprotein fractions from all hyperlipemic sera tested had HI activity (titers ranging from 1:16 to 1:1,024) which, in the majority of cases, was not eliminated after heparin-MnCl(2) treatment. In every case, treatment with SPS-CaCl(2) removed this nonspecific activity completely. Since hyperlipemic sera may occasionally be encountered in routine rubella HI antibody testing, we recommend the use of SPS-CaCl(2) rather than heparin-MnCl(2) for pretreatment of sera.