Evoked visual potentials were studied in 12 subjects diagnosed or suspected of having MS. A comparison was made between these subjects and results obtained with single eye stimulation using a flash of light (101 subjects) or by using a reversible checker-board pattern (58 subjects). Reversible checker-board pattern simulation showed pathological unilateral or bilateral increase in the latency of the main positive spike in the response of 80% of the verified MS cases and 50% of the probable or possible MS cases. These percentages were 25% lower when flashes of light were used as stimulation. 58% of those patients who have never presented ophtalmological symptoms have abnormal latency in responses invoked by the reversible checker-board pattern, as opposed to 20% using flashes of light for stimulation. Study of E.V.P. in normal subjects indicates that the better results obtained with reversible checker-board stimulation can be attributed to greater reproductibility of the response. Analysis of the overall morphology of the responses significantly increases E.V.P. examination efficiency when using flashes of light as stimulation. However, in our series, it does not change the results obtained by only calculating the latency of response to reversal of a checker-board pattern.