Hue reversal in the Fechner-Benham color effect following white light adaptation.

The Fechner-Benham color phenomenon is a percep tual effect in which color is generated by intermittent presentations of patterned white light. The most common method for producing the colors involves rotating a black and white patterned disc at speeds well below fusion. This method was first introduced by Fechner (1838), and since then, a multitude of disc designs have appeared in the literature. The best among them is the design developed by Benham (1894), shown in Fig. 1. Benham’s design represented a major advance in the study of the effect, since it was the first design to segregate spatially the various colors produced. When the disc is viewed under incandescent illumination (Illuminant A) and rotated counterclockwise at about 8 Hz, four concentric rings of different colors are produced: ring 1, the outermost, appears red; ring 2 appears yellow-green; ring 3 appears either dark green or grey; and ring 4 appears either dark blue or black. Clockwise rotation reverses the sequence of colors. The colors appear much weaker when viewed under bluish-white light (Illuminant C), the rings appearing primarily grey (Verriest and Seki, 1964). The purpose of this report is to relate what we believe to be an original observation concerning this effect. The observation was made in the course of examining the possibility of rod involvement in the production of the colors (a hypothesis which incidentally was excluded). To test this hypothesis, we employed a simple two-step procedure. Observers were first asked to fixate monocularly the center of a white (50,000 cd/m’, Illuminant A; or 11,000 cd/m’, Iliuminant C), 10” circular adapting field, for 1 min. They were then instructed to shift their gaze from the adapting field toward a rotating Benham disc viewed under Illuminant A (6OOcd/m’) and asked to give a verbal description of each of the disc’s four rings. The following sequence of changes in ring appearance under Illuminant A is an exact description of color seen by the senior author with a similar sequence observed by five other color-normal observers. For the initial 10sec of viewing following adaptation to either Illuminant A or C, all rings were reported as achromatic, and a pink afterimage due to the adapting light was seen covering the surface of the disc. Following this initial period, color was observed in rings 1 and 2. But, unexpectedly, the colors seen were the complements of those normally seen at those postions. Ring 1 which is normally seen as red was reported as green, and ring 2 which is normally seen as green was seen as red. Rings 3 and 4, the two innermost, still remained achromatic. The pink afterimage, though very much weaker, was still visible. This color reversal interval lasted for about 50sec in ring 2 before it returned to green. During the next 30 set interval, rings 1 and 2 were both seen as green. Following this period, ring 1 reverted to its usual red hue. Hue shifts in rings 3 and 4 were never reported. After 2.5 min. the disc looked the same with the adapted eye as it did with the nonadapted eye. Moreover, during the pre-adaptation stage of the experiment, all observers were asked to match Munsell chips (viewed under Jlluminant A) to the perceived hues of rings 1 and 2 in a grey surround of equivalent Talbot luminance to the rotating disc. Following adaptation, no hue reversals in steadily viewed Munsell chips were reported. Further, viewing the Munsell chips under flickering light (507; duty cycle) did not induce color reversal following adap