Visually evoked potentials to small patterned stimuli were recorded in an occipital row of five electrodes. The location of the flashed checkerboard stimuli of three different spatial frequencies varied in five steps from 5 degrees on the left to 5 degrees on the right hemiretina. Repeated recordings were obtained from seven healthy subjects and one Multiple Sclerosis patient. Stimulus location, electrode position and their interaction had stronger effects on the components derived from a principal components analysis than the spatial frequency. Components most significantly related to these spatial frequency or to an interaction between spatial frequency and other experimental conditions showed later peak latencies than those related to electrode position or stimulus location. This finding suggests a later or more sustained processing of pattern information. Discriminant analyses using the patient's component scores, separate principal components analysis, and the analysis of the patient's residual components all suggested major differences between the patient and the normals.