Brain potentials (lambda responses) associated with eye-fixation pauses were measured under four visual stimulus conditions. Ten students were presented monotonous gray paper (MT), wide stripes (WS), narrow stripes (NS), or gradational stripes (GS). EEGs and EOGs were recorded when the subject moved his eyes onto each stimulus. EEGs time-locked to offset of saccades were averaged to obtain the lambda response. The amplitude of the positive component (lambda response) with a latency of about 100 ms was greatest in the NS condition, and decreased in the order of WS, GS, and MT. The negative component with a latency of about 180 ms also showed variations similar to the lambda response. These results suggested that contours of the visual stimulus are necessary for appearance of the lambda response, like the visual evoked potential.