Journal of Vision, 5. 2005. [VSS 2005; Sarasota, FL, USA.]

When a large and a small concentric square alternate in succession with a blank separating the two, observers often report seeing one or more distinct illusary squares (or ÒghostsÓ) in the blank field. To explore this ÒGhost IllusionÓ, squares with black edges and white backgrounds were used, ranging between xx and yy degrees in size. Each square was displayed for 80 ms; the two displays alternated 5 times, with a white screen (duration 125 ms) between each display. Observers were informed that more than one square size would appear. They were instructed to count the total number of squares differing in size that appeared, and then respond using the numbered keys. Several experiments were carried out, with twelve observers in each, and 36 trials per observer. The first three experiments tested a range of size ratios (2:1-4:1). The strength of illusion decreased only slightly as the size ratio increased. Two, three and four squares were perceived in 31%, 49% and 16% of trials for size ratio 2:1, in 37%, 39%, and 13% of trials for ratio 3:1, and in 32%, 33%, 18% of trials for ratio 4:1, respectively. Similar results were obtained when the thickness of the edges was tripled for ratio 2:1 and when the colors of the edges were distinct (red and green) for ratio 4:1. These results therefore show that the strength of the illusion persists across size ratio, edge thickness, and color.

Support provided by the Natural Sciences and Engineering Research Council of Canada.