Flicker electroretinograms (FERGs) of 30Hz were recorded for 46 eyes (23 cases) of low vision, including 38 eyes with retinal diseases and 8 eyes with optic neuropathy, and for 39 normal eyes (33 cases). The recorded wave forms were analysed by two methods: (1) to measure the peak to peak amplitude and pseudophase directly, and (2) to measure the amplitude and phase of fundamental response component (30Hz) by discrete Fourier transform (DFT). The abnormality ratios between these two methods were compared in low vision patients. All the eyes with optic neuropathy showed normal flicker ERG, the abnormality ratios of directly measured amplitude and the amplitude by DFT were 60.9% in 28 eyes and 71.7% in 33 eyes with retinal diseases respectively, the difference being no statistically significant (P > 0.05), while the abnormality ratios of directly measured pseudo-phase and the phase by DFT were 60.9% in 28 eyes and 82.6% in 38 eyes with retinal diseases respectively, the difference being statistically significant (P < 0.05). These results suggest that FERGs are an objective way to test the retinal function, and DFT for flicker ERGs are more useful in quantitative evaluation of the retinal function.