The purpose of this work is to investigate the use of different forms of visual evoked potentials (VEPs) to measure color discrimination thresholds and to plot color discrimination ellipses (MacAdam, 1942). Five normal trichromats (24.5 +/- 2.6 years-old) were monocularly tested. Stimuli consisted of sinusoidal isoluminant chromatic gratings made from chromaticity pairs located along four different color directions radiating from one reference point of the CIE 1976 chromaticity diagram (u' = 0.225; v' = 0.415). Heterochromatic flicker photometry (HFP) was used to obtain the isoluminance condition for every subject and for all chromaticity pairs. VEPs were elicited using two cycles per degree grating stimuli at three different temporal configurations: transient, onset (300 ms)/offset (700 ms), 1 Hz fundamental frequency; steady-state, onset (50 ms)/offset (50 ms), 10 Hz fundamental frequency; and steady-state pattern reversal at 5 Hz fundamental frequency (10 Hz phase reversal). VEP amplitude was measured using transient VEP N1-P1 components and steady state VEP first (10 Hz) and second (20 Hz) harmonics. VEP amplitude was plotted as a function of chromatic distance in the CIE 1976 color space and the data points were extrapolated to zero amplitude level to obtain chromatic discrimination thresholds. The results were compared with psychophysical measurements performed using the same stimulus configurations and with the pseudoisochromatic method of Mollon-Reffin (one-way ANOVA). For all subjects and all stimulation methods, the ellipses showed small sizes, low ellipticities, and were vertically oriented. Despite some consistent differences in the results obtained with different procedures, there was no statistical difference between ellipses obtained electrophysiologically and psychophysically. For steady state VEPs, ellipses obtained from second harmonic amplitudes were larger and more elongated in the tritan direction than those obtained with first harmonic amplitudes.