The pattern electroretinogram (PERG) is probably generated by the ganglion cell layer of the retina whose axons constitute the conductive fibres of the optic nerves. A study was undertaken of the PERG in a group of patients with clinically definite multiple sclerosis who had suffered optic nerve demyelination to assess the value of recording the PERG as a method for electrophysiological determination of optic nerve dysfunction. Pattern visual evoked potentials (PVEP) were simultaneously recorded, each eye being stimulated separately, and four different check sizes were used. Although individual patients showed PERG abnormalities the results as a group showed no significant difference from the normal population so far as latency or amplitude of the PERG was concerned. Thus we conclude that the PERG, while an interesting measurement, does not satisfactorily reflect optic nerve dysfunction. Further, our results indicate that it is not such a useful measure as the pattern visual evoked response for detecting optic nerve disease.