No abstract

Subjects (groups of age 6, 8, 10, 12, and 21 years, with 60 or more persons per group) made comparative size judgments between a single circle and a standard circle that was flanked by four context circles. The context circles varied in diameter, proximity to the standard circle, and amount of circumference displayed. (When circumferences are complete, the display is the usual Ebbinghaus configuration. A variant consisted of the Ebbinghaus display with the outer three quarters of each context-circle circumference removed.) A contour-plus-context theory accounts for the data: Contours attract at every age level, and context (size of flanking circles relative to the standard circle) leads to constrast in judged size beyond age 6. Contour and context effects decreased with decreasing proximity between context circles and the standard circle.

Variants ofthe Delboeuf (concentric circle) display and the Ebbinghaus display (four large nonconcentric inducing circles) were produced by systematically eliminating fractions of the circumferences of the outer inducing circles. For the Ebbinghaus display, various amounts of the inner arcs and/or outer arcs were preserved; a series of dots was sometimes substituted for missing contours. Results are described by the following statements: Contours attract. Larger nonconcentric circles induce size contrast. Rotating an Ebbinghaus variant by 45°alters size misjudgments. The larger the horizontal (or vertical) extent of the inducing configuration, the smaller the judged size of the test (center) circle. A figural aftereffect of the inducing configuration acts to reduce the perceptual size of the comparison circle when eye movements are permitted between test and comparison circles.The term illusion, because of its connotations of magic and deception, neither of which is germane to the topic at hand, shall be restricted to the title. Informativeness requires its inclusion in the title. At issue are explanations of the Delboeuf anomaly, size misjudgments of a circle in the presence of a concentric circle, and the related Ebbinghaus (Titchener circles) anomaly, which depends upon the presence of nearby nonconcentric circles. The underlying problem is to understand the processes underlying misperceived circle size. The present investigation relies upon displays that are sometimes only distant cousins to Delboeuf and Ebbinghaus figures.The research was designed to explore elements of a theory proposed on the basis of relatively modest evidence (Cooper & Weintraub, 1970;Weintraub & Cooper, 1972) but more recently buttressed by extensive evidence (Weintraub, 1979). In barest outline, the contour/circle-context model asserts that contours attract and that context supplied by the larger or smaller size of nearby nonconcentric circles leads to size contrast. Contour attraction, then, predicts that exterior contours produce overestimation of Experiment 2 was based on research by the second author for his senior honors thesis in the Psychology Department, University of Michigan.