We evaluated a new agarose-gel-electrophoretic procedure (Corning) (I) for separating and quantitating of high-density lipoprotein cholesterol (HDLC), comparing it with the modified Lipid Research Clinics (LRC) procedure (heparin 183 kilounits/L, MnCl2 92 mmol/L) (II). Method I was insensitive to an HDLC concentration of 50 mg/L, but gave a linear dose-response curve between 130 and 1200 mg/L. Method II is sensitive to 50 mg/L and linear from 50--1200 mg/L. The within-plate CV for the Corning method varied from 26.2% for an HDLC of 168 mg/L to 6.8% for 580 mg/L. Within-day between-plate CV for the Corning method ranged from 22.1% at 155 mg/L to 8.0% at 651 mg/L, compared to 3.0 and 0.8% for the modified LRC procedure. Between-day CV for method I was 20, 12.6, 4.3, and 3.5% for HDLC concentrations of 175, 435, 542, and 678 mg/L, respectively; for method II it was 14, 5, 3.5, and 2.6%, respectively. Analysis of HDLC in 100 patients by both procedures showed mean HDLC values to be significantly lower (mean + SD, 27.8 +/- 1.7 mg/L; p less than 0.001) by method I. In 46 patients with HDLC less than 450 mg/L, this difference was accentuated (mean + SD = 40.5 +/- 2.6 mg/L) and clinically significant. Electrophoretic methods offer a promising further alternative method for HDLC separation and quantitation, but the negative bias, present limited sensitivity, and lack of precision at less than 450 mg/L indicate that they are not yet optimal for routine clinical use for patients with values less than 450 mg/L.