Abstract
When the interior segments of the inducers of a Kanisza square are colored, the entire square appears to be of the same hue but lower saturation. Conventional methods of studying this effect (e.g., reporting its presence or subjecting ranking of its strength) do not specify its perceived saturation quantitatively. The role of the chromatic saturation and luminance contrast of the inducing elements was investigated by hue cancellation in a two-alternative forced-choice task, with variable saturations of the cancellation stimulus. At low chromatic saturations, increasing the saturation of the inducers leads to increases in the saturation required to cancel the spread color. However, this trend reaches an asymptote at moderate-to-high saturations, beyond which further increases in the chromatic saturation of the inducers do not affect hue cancellation.
Changes in the luminance contrast of the display resulted in relatively smaller changes in the magnitude of color spreading. These do not appear to interact with the changes caused by chromatic saturation. Increases in luminance contrast result in increases in the saturation of the hue cancellation stimulus.
The results are discussed in relation to models of color spreading. The findings provide general support for the concept of two independent mechanisms, one responsible for boundary formation and the other for the spreading of the color.