This research investigates the method used for grouping each subject's visual scanning patterns and contains a procedure to match the criterion position in experiments and analyses.　Nine truck drivers were fitted with an eye-mark recorder and viewed 14 different traffic scenes recorded on videotape.　This paper addresses visual scanning patterns of two basic scenes, a straight road and a curved road.　These scenes were divided into four categories (straight road, approaching curve, curve negotiation, and exiting curve), with each category analyzed separately.　Each subject's visual scanning patterns were calculated, figured, and then grouped into a total figure in each category.　The results were as follows.　(1) All subjects gazed at the center of the front screen within a visual angle of 6° for a straight road scene (60% to 70% of the gazing time).　(2) The total figure of visual scanning patterns shifted left by 3° to 5° (visual angle) from the center point for an approaching left-curve scene.　(3) The total figure shifted more to the left by 9° to 11° from the center point for a left-curve negotiation scene.　Pairs of each category were tested by t test, and all pairs exhibited significant differences.　These results were consistent with the results of an on-road field study conducted by other researchers.