A more extensive list of facts and inferences:
- The illusion occurs over a large range of sizes and bar widths
- The illusion is reduced by use of curvy lines and discontinuities
- Using configuarations that increase center surround antagonism
actually decreases the effect
- The effect is enhanced with luminance differences of bars
but occurs only when lesser contrast bars are in front
- The actual spatial arrangement of the retinal ganglion cell
receptive fields is in conflict with the theory
- The illusion persists with contrast reversal
Therefore
the theory based on ON ganglion cell activity is invalid
- Illusion not seen at isoluminance
- Illusion is seen with brief presentation and is not altered
by light and dark preadaptation
Therefore
simultaneous luminance contrast is essential
Therefore
eye tremor and afterimages are not a factor
- The effect is strong under both scotopic and photopic conditions
Therefore
effect not due to rod/cone interaction
For earlier studies examining the Hermann grid and the possible underlying
mechanisms of the illusion, see Lothar Spillman (1971) Pflugers Archiv,
326, 281-299 and Jeremy M. Wolfe (1984) Perception, 13,
33-40.