Normal corneal electroretinograms (ERG) are analyzed in the frequency domain using the fast Fourier transform (FFT) and linear prediction (LP) methods. Four dominant frequencies at 18, 79, 126, and 159 Hz are found in the dark-adapted state. Light adaptation shifts the low frequency to higher frequency and the mid- and the two high-frequency components to lower frequencies. The relative amplitude of the high-frequency component resulting from the oscillatory potentials is quantified. It is shown that frequency-domain features are of a smaller variability than time-domain components, and can be extracted even from a noisy surface ERG.