This investigation compared four methods used to control eye-blink artifact in visual evoked response (VER) recordings using the augmenting-reducing paradigm: rejecting trials containing blinks, presenting the stimuli through closed eyelids, and two different techniques for subtracting a proportion of the averaged electro-oculogram (FOG) from the averaged VER. The two subtraction procedures differed in the way in which they estimated the amount of EOG to be subtracted and in the case with which they could be applied. Twenty normal, female subjects were exposed to a series of four different intensity light flashes twice with eyes open and once with eyes closed. Blinking was monitored by recording the EOG and EEG was recorded from three midline electrodes. The results indicated that the two subtraction techniques were very similar. They effectively removed ocular artifact and produced corrected VERs that showed high retest stability. Presenting the flashes through closed lids yielded idiosyncratic data, and rejecting trials contaminated by blinks was found to have limited utility with the VER paradigm used in the present study.