Open a music box and the sound is coming off a comb: a row of steel teeth of graded length, each one plucked in turn by pins set into a slowly turning cylinder. It is one of those mechanisms that looks obvious for about four seconds. Short teeth make high notes, long teeth make low notes, done. Then you ask how low, and the obvious picture quietly falls apart.
Each tooth is a cantilever, clamped at one end and free at the other, and the pitch of a cantilever is set by how stiffly it springs back divided by how much mass has to move, under a square root. The length sits in that stiffness as an inverse cube, which means halving a tooth makes it eight times stiffer against a quarter of the mass, and the pitch climbs as roughly one over the length squared. That is why the treble end of a comb is a little picket fence of stubby teeth. It is also why the bass is a problem.
To drop an octave you need to quarter the frequency, and by length alone that means doubling the tooth. The next octave down doubles it again. A proper bass note would want a tooth far too long to fit the movement and far too floppy to pluck cleanly, so somewhere below the middle of the comb, lengthening stops being a usable lever. The craftsmen who settled the design two hundred years ago reached for a different one: they loaded the tip. A small weight soldered to the underside of a bass tooth adds mass without adding length, and the same short tooth sings lower. The lowest teeth on an old comb carry a little block of lead on the end for exactly this reason.
The part I find genuinely lovely is why lead. The teeth are hardened spring steel, and their temper, the heat treatment that gives them their ring, is destroyed above roughly 400 degrees Celsius. Lead melts at 327. So you can solder a lead weight onto a finished, tuned, tempered tooth without ever heating the steel to the point where it would go soft and lose its voice. The metal is chosen for the one property that lets you attach it without wrecking the thing you are tuning. It is a solution shaped entirely by a constraint most people never see. (Modern combs, lead now being unwelcome, reach the same notes by pinning on a tiny brass weight instead, or by thinning the root of the tooth to soften its spring. The constraint moved, so the craft moved with it.)
There is one more thing hiding in the sound. A plucked string rings in a neat harmonic ladder, two and three and four times its fundamental, which is what makes it read as a clear musical pitch. A struck bar does not. Its higher modes ring at about 6.3 and 17.5 times the fundamental, ratios that are not musical multiples of anything, and those off-ladder partials are exactly the cool, glassy shimmer sitting on top of every music box note. You are not hearing a string. You are hearing the shape of a bending bar, and the bar is honest about being a bar.
I built a small interactive to make all of this audible. You get one tooth. Drag its length and watch the pitch rocket; try to reach the bass by length alone and run out of comb; then slide a lead weight onto the tip and hear it fall more than an octave and a half without growing a millimetre. The frequencies are computed live from beam theory, with the real inharmonic overtones baked into the synthesiser, so the thing you are hearing is the physics and not a recording.
What I like about this object is how much of it survives. You can decarbonise the factory, swap the lead for brass, cut the comb by CNC instead of by hand, and the governing fact is still the one a cantilever had in 1810: pitch is stiffness over mass, and when you run out of room to change the stiffness, you change the mass. The lead in the bass is a two-hundred-year-old answer to a question the physics has been asking the whole time.