The Water Joins the Instrument

Imagine tuning an instrument by pouring a little more of your drink into it. The material stays the same. The rim stays the same. Yet the note goes down.

A singing wineglass makes the trick look almost too simple: rub the rim, add water, find another note. The explanation I wanted was not just that water makes the glass “heavier.” Heavier in what motion? And why should the same amount added near the bottom count differently from water added near the rim?

The answer begins with a glass doing something its everyday solidity conceals.

A circle borrows two ellipses

In the simplest wineglass mode, the rim does not swell outward everywhere at once. It bends toward an ellipse. Two opposite sides move out while the other two move in. The rim passes back through its circular shape and bends into the perpendicular ellipse. There are four places around it with maximum motion, separated by four nodes with very little radial movement. Those are nodes of radial motion; the glass also moves tangentially around the rim. Farther down the bowl, the motion diminishes toward the supporting base.

That is the picture A. P. French develops in his 1983 paper, In Vino Veritas. It lets the glass bend while keeping the rim's circumference approximately unchanged, rather than demanding that the material repeatedly stretch and compress around the whole circle. The tiny deformation is enough; the glass need not visibly wobble like a rubber cup. [1, §II]

The finger supplies energy, not a timetable for every vibration. In the stick-slip description, contact first drags the rim and then releases it, sustaining motion against losses. Denis Terwagne and John Bush explain the same process for a mallet rubbing a singing bowl. Their recordings distinguish the several natural modes excited by a strike from the fundamental and harmonics prominent during rubbing. Slowly circling a rim is not the same thing as slowly vibrating it. [2, §2]

Once I had that moving shape in mind, the water stopped looking like something merely occupying space inside the instrument.

What the glass has to carry

The bending wall drives liquid motion. For the same motion of the glass, the glass-and-water system therefore carries more kinetic energy than the empty glass. In French's simplified calculation, the elastic restoring energy for that prescribed glass deformation stays the same. More inertia, with the same stiffness, means a lower natural frequency. [1, §VI]

This is not energy being used up and disappearing into the water on every cycle. Kinetic energy is stored motion, available to be exchanged with elastic energy. Dissipation is another question. Water joins the moving system before we have said anything about how rapidly that system loses energy.

There is a useful numerical warning here. In an ideal oscillator, doubling the effective moving mass does not halve the frequency: it multiplies it by one divided by the square root of two, about 0.707. “Twice as heavy, half as high” is the wrong intuition.

French's measurements give a less hypothetical example. One glass had an empty frequency of 4,600 Hz. With the liquid surface just 0.1 centimetre below the rim, the reported frequency was 3,100 Hz—about two-thirds of the empty value. Under the unchanged-stiffness interpretation, that corresponds to roughly 2.20 times the original effective inertia. It does not mean the glass and its contents weighed 2.20 times as much. The calculation concerns participation in this particular motion, not the reading on a kitchen scale. [1, Tables I–II; calculated ratio]

Where the liquid sits matters because the wall does not move equally at every height. Near the relatively still base, it drives less liquid motion; toward the lively rim, added liquid can contribute much more.

French's approximate cylindrical model captures this with a fourth-power dependence on relative fill height in the added-inertia term. At half the effective height, that term is only one-sixteenth of its full-height value. That is not one-sixteenth of the pitch change, and it is not a tuning recipe for every goblet: French had to infer an effective vibrating height for glasses whose walls were not uniform. But the model gives a reason equal steps upward through the glass need not produce equal steps downward in frequency. [1, Eqs. 17–21]

A bowl is a useful second instrument

Terwagne and Bush studied four antique metal singing bowls. Their water-height series for one bowl, Tibet 1, showed its fundamental frequency falling as the fill increased. They compared the measurements with the same kind of height-dependent loading relation. [2, Table 1 and Fig. 5(a)]

This is a different team's experiment on a different kind of vessel—not a second independent invention of the theory. Their model explicitly builds on French. Nor does one bowl's fill series establish the exact tuning curve of every wineglass. What carries across is the physical idea: an elastic vessel can make liquid participate in its vibration, and that participation changes the note.

It also helps explain why “more water makes it quieter” would be a separate claim. A natural frequency does not specify how hard an instrument is driven, how much motion survives its losses, or how efficiently that motion sends sound into the air. A lower note alone cannot answer all three questions. I would not turn a frequency measurement into a loudness law.

The distinction I want to keep is simple: water is not only something that can interfere with an instrument. Here it becomes part of the instrument's moving body.

The next experiment worth understanding would measure both the wall's motion and the decay of its vibration in the same glass as the water rises. How much of the changing response comes from carrying more liquid, and how much from losing energy differently? A note tells us something real. It does not tell us everything the instrument is doing.

Sources and scope

1. A. P. French, “In Vino Veritas: A study of wineglass acoustics,” American Journal of Physics 51, 688 (1983), consulted institutional PDF, DOI. Complete article read through OCR, including methods, tables and captions; its nonuniform-glass and higher-mode uncertainties remain relevant.
2. Denis Terwagne and John W. M. Bush, “Tibetan singing bowls,” Nonlinearity 24, R51–R66 (2011), author-hosted PDF, DOI. This account uses the abstract and complete Acoustics section, including its measurement descriptions, tables and captions—not the later fluid-wave study or supplementary movies.

The opening is imagined. I did not play, record or listen to a glass for this piece. The ratios are illustrative calculations, not a new experiment, plot digitization or raw-data reanalysis.