We examined three immunoassay techniques for measuring apolipoprotein B in serum and major lipoprotein density fractions from normolipidemic and hyperlipoproteinemic persons, comparing values by electroimmunoassay, radioimmunoassay, and radial immunodiffusion assay with those determined gravimetrically. Electroimmunoassay is faster and simpler than radioimmunoassay, and equally precise (within- and between-assay coefficients of variation for both were 5 and 7%, respectively). All the immunoassays gave results that agreed with those by gravimetry for normolipidemic sera and the corresponding lipoprotein density fractions, but only electroimmunoassay results agreed with those by gravimetry for apolipoprotein B in lipoproteins of d less than 1.019 g/ml isolated from hypertriglyceridemic patients. Concentrations of apolipoprotein B in plasma, determined by electroimmunoassay in a population of normal persons and patients with primary hyperlipoproteinemias, were: normals, 980 +/- 200; type I, 700 +/- 160; type IIa, 2000 +/- 260; type IIb, 2180 +/- 300; type III, 1300 +/- 340; type IV, 1470 +/- 400; and type V, 1550 +/- 390 mg/liter (mean +/- SD). Lipoprotein density fractions from the hyperlipoproteinemic patients each had a characteristic distribution of free and associated forms of lipoprotein family B. The absolute concentration and distribution of apolipoprotein B between the free and associated forms of lipoprotein B may represent a useful indicator of the underlying biochemical defect(s).