THE RANGE OF VISUAL SEARCH

The aim was to study the process of visual search in its early phases. Individual human subjects searched in a projected matrix of elements for one element unlike the rest; e.g., for a triangle in a matrix otherwise composed of circles. In the method of lasting exposure, the matrix was exposed until the subject responded, and the dependent variable was the latency of the response. In the method of brief exposure, the exposure time was limited, and the dependent variables were the percentage of positive responses and the latency of the positive responses. The graph begins at nearly zero slope, and usually shows a small discontinuity or a sigmoid transition leading to slightly higher latencies. This triangle locates the critical number (CN): the number of elements at which it occurs. The critical number varies considerably with the type of discrimination required, and with the stimulus difference between the critical and background elements. The critical number represents some discriminatory limit or limits; it may be a limit of area rather than of number. In dealing with large matrices, the subject apparently searches rapidly in a region around the fixation point (the initial sub-matrix). By definition, this has an area equal to that covered by the critical number in the matrix. The interpretation of the critical number and the initial sub-matrix is partly in terms of saccadic eye movements, though none have been photographed. The region of fast search may have at least an approximate shape. One experiment, using the method of brief exposures, indicated the shape to be ovaloid, with most of its area lying above the fixation point.