Estimates of serum or plasma high-density lipoprotein (HDL) concentrations are usually accomplished by measuring the residual cholesterol after polyanionic precipitation of the very low and low-density lipoproteins. A comparison was made of the HDL cholesterol values obtained using both heparin--MnCl2 and dextran sulfate--Mg2+ as precipitating agents. Enzymatic cholesterol determinations showed that the dextran sulfate--Mg2+ method led to gross underestimation of the serum HDL cholesterol concentration (approximately -25%), a difference that was even greater when ultracentrifugally isolated HDL concentrations were assayed rather than serum. Immunochemical determinations of the A-I and B apolipoproteins demonstrated that the discrepancy was attributable neither to incomplete precipitation of the lower density lipoproteins by heparin--MnCl2 nor to precipitation of HDL by dextran sulfate--Mg2+. Incubation of sera with reagents at 37 C, however, minimized or eliminated the differences in results. These findings indicate that temperature is a critical factor when enzymatic quantification of HDL cholesterol is performed after dextran sulfate--Mg2+ precipitation.