Pattern reversal visually evoked responses (VER's) at a frequency of 7 Hz were recorded from 10 subjects. Peak-to-trough amplitudes were measured at centrally fixated field sizes from 1 to 8 degrees for check sizes from 2.5 to 30 min. Check size for maximum amplitude decreased as field size was reduced. The amplitude--check size relation was not necessarily linear across its full range. When using an extrapolation technique for estimating visual acuity, a curvilinear regression may be more appropriate than the assumption of a linear regression.