We compared the best-corrected Snellen acuity (SA) and the pattern reversal visual-evoked response (PVER) acuity in normal subjects and patients. Forty-two eyes of 42 normal subjects were controls; 457 eyes of 329 patients comprised the patient group. A steady-state stimulus with five check sizes ranging from 160 to 10 min in 1.0-octave steps was used. The PVER acuity was derived from the best-fit linear function relating the amplitude to the log-adjusted check size. Three intercepts of 0, 1 and 2 microV were used in both groups, and the PVER acuities were called P0, P1 and P2. The SAs in normal subjects ranged from 20/15 to 20/20 (mean, 20/18.3) and in patients from 20/15 to 20/1,600 (mean, 20/56.9). In normals, the P0 showed the best agreement with the SA (mean acuity difference, +0.34 octave). The SA and P0 agreed within +/- 2.0 octaves in 33/42 (78.6%) eyes. In patients, the P0 also showed the best agreement with the SA; 306/457 (67.0%) eyes showed an acuity difference within +/- 2.0 octaves. Unlike normals, 83/457 (18.2%) eyes showed an acuity difference > -3.0 octaves. These eyes mostly had optic nerve disease with a flattened PVER amplitude-check size function curve. The P0 seems to correlate better with SA than P1 and P2, but this analytical method may be less effective in the presence of certain pathologic conditions.