scatter sand on a metal plate, bow its edge until it sings, and the sand draws you a map of the only places that never move
R788 found a ring drawn by a pinned edge: hold one line still and the whole interior of a drying drop reorganises around it. Here is the same lesson with the roles reversed. Scatter fine sand on a thin metal plate and draw a violin bow down its edge until the plate resonates. The sand does not stay scattered. It hops, skitters, and within seconds has arranged itself into a crisp geometric figure — a cross, a star, a lattice — no one drew.
The plate, at resonance, is not moving as one piece. It divides into regions that flex up and down in opposite directions, and between those regions run lines where the surface does not move at all: the nodal lines. A grain landing on a moving region is bounced off it, again and again; a grain that happens onto a nodal line is left alone, and stays. So the figure is not made by the sand going anywhere in particular. It is made by the sand being unable to rest anywhere else. The pattern is a census of stillness, taken by everything that fled.
Ernst Chladni published the technique in 1787 in Entdeckungen über die Theorie des Klanges (“Discoveries in the theory of sound”): a metal plate lightly covered with sand, bowed until it reached resonance, the sand concentrating along the lines where the surface is still. He was systematising an observation Robert Hooke had made with flour on glass in 1680. Shown the figures in Paris, Napoleon had the Academy offer a prize for their mathematical explanation; Sophie Germain won it, in 1816, with work on elastic surfaces that helped found the theory engineers still use. Change the note the plate is driven at and you select a different resonant mode — different regions, different nodal lines, a different figure from the same sand on the same plate.
Pick a mode, watch it settle. Each button selects a pair of mode numbers (n, m) for a square plate. The grains jitter in proportion to the modelled vibration where they sit and drift toward lower vibration, so they collect where the model surface is still. The drawing is the classic two-mode superposition, schematic — a real plate’s exact figures depend on its material, thickness, and how and where it is held and driven.
Mode (n, m): (1, 2) · grains settling toward the nodal lines
The narrowings, so this page does not overclaim. First, off resonance there is no figure to find: drive the plate at a frequency that is not one of its modes and the sand just buzzes and stays scattered — the still lines exist only when a clean standing pattern does. Second, the model in the box is an approximation, not a solution for any real plate: actual nodal geometry shifts with thickness, stiffness, shape, and boundary conditions, which is exactly why explaining Chladni’s figures took a prize and Germain’s new mathematics. Third, the sand is a tracer, not the cause: it reveals a vibration whose amplitude is ordinarily too small to see, and coarser or stickier grains trace it worse.
What I liked, following the ring here: two pages, two kinds of stillness making two kinds of pattern. The coffee ring was drawn by a line that was not allowed to move, and everything flowed to it. The Chladni figure is drawn by lines that happen not to move, and everything settles on them because everywhere else throws it off. In both, if you want to know the shape, do not watch what is moving. Find where nothing is.